submission.csv

candidate_id,rank,score,reasoning
CAND_0081846,1,0.986854,"Lead AI Engineer with 6.7 years at Razorpay, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Razorpay.; Designed and built a significant technical initiative at Paytm.; Designed and built a production system at Paytm. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0046064,2,0.986709,"Senior NLP Engineer (Salesforce, 8.9 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Verloop.io.; Designed and built a significant technical initiative at Amazon.; Designed and built a end-to-end production system at Amazon. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions."
CAND_0080766,3,0.979639,"8.8-year Staff Machine Learning Engineer at Salesforce, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Designed and built a dense vector retrieval system leveraging ann (ann) at Aganitha.; Designed and built a significant technical initiative at Aganitha.; Designed and built a end-to-end production system at Aganitha. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and has the architected/owned-system track record a founding role requires. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0071974,4,0.979472,"7.8-year Senior AI Engineer at Netflix, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Designed and built a dense vector retrieval system leveraging pinecone (Pinecone) at Netflix.; Designed and built a significant technical initiative at Netflix.; Designed and built a end-to-end production system at Netflix. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and has the architected/owned-system track record a founding role requires. Consideration: could strengthen the retrieval infrastructure depth the role demands."
CAND_0030953,5,0.973796,"Search Engineer with 7.8 years at Nykaa, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Vedantu.; Designed and built a significant technical initiative at Nykaa.; Designed and built a production system at Nykaa. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: evidence of A/B testing and offline-online evaluation could be stronger."
CAND_0017960,6,0.970946,"7.7-year Recommendation Systems Engineer (Nykaa) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Sarvam AI.; Designed and built a significant technical initiative at Nykaa.; Designed and built a production system at Nykaa. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0039383,7,0.969385,"Applied ML Engineer at Meesho (7.1 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Swiggy.; Designed and built a significant technical initiative at Meesho.; Designed and built a production system at Meesho. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0053695,8,0.967435,"Recommendation Systems Engineer with 5.8 years at Meesho, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Niramai.; Designed and built a significant technical initiative at Genpact AI.; Designed and built a production system at Meesho. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0075574,9,0.966902,Machine Learning Engineer at Haptik (5.7 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Genpact AI.; Designed and built a significant technical initiative at Haptik.; Designed and built a production system at Observe.AI. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar.
CAND_0064326,10,0.963989,"Search Engineer with 7.6 years at Sarvam AI, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Freshworks.; Designed and built a significant technical initiative at Sarvam AI.; Designed and built a production system at Aganitha. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: the profile would benefit from more quantified outcome evidence."
CAND_0044883,11,0.963599,"6.3-year AI Engineer (Mad Street Den) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Meesho.; Designed and built a significant technical initiative at Paytm.; Designed and built a production system at Paytm. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the architected/owned-system track record a founding role requires. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0011432,12,0.962827,"Senior Data Scientist (Amazon, 7.6 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Krutrim.; Designed and built a significant technical initiative at Amazon.; Designed and built a production system at Amazon. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the architected/owned-system track record a founding role requires. Consideration: not active on the platform for 93 days."
CAND_0018722,13,0.961481,"6.6-year Recommendation Systems Engineer (Saarthi.ai) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Saarthi.ai.; Designed and built a significant technical initiative at Unacademy.; Led development of an evaluation framework at Saarthi.ai. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0044222,14,0.960848,"7.7-year AI Engineer (PolicyBazaar) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at PolicyBazaar.; Designed and built a significant technical initiative at InMobi.; Designed and built a production system at InMobi. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and has the architected/owned-system track record a founding role requires. Consideration: could strengthen the retrieval infrastructure depth the role demands."
CAND_0068811,15,0.953724,Applied ML Engineer at Freshworks (8.0 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Yellow.ai.; Designed and built a significant technical initiative at Meesho.; Led development of a end-to-end production system at Salesforce. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: the profile would benefit from more quantified outcome evidence.
CAND_0005649,16,0.951282,"Senior Data Scientist with 7.4 years at Sarvam AI, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Sarvam AI.; Designed and built a significant technical initiative at Glance.; Designed and built a production system at Glance. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0083307,17,0.942921,"7.8-year Search Engineer at CRED, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Ola.; Designed and built a significant technical initiative at Netflix.; Led development of an evaluation framework at Ola. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: notice period of 120 days is above the JD's sub-30-day preference."
CAND_0069905,18,0.942521,"6.6-year Applied ML Engineer (Sarvam AI) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Sarvam AI.; Led development of an evaluation framework at Sarvam AI; Led development of pre-2022 search/ranking infrastructure (elasticsearch, bm25) at Observe.AI. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and brings classical IR depth AND system-ownership evidence — the JD's stated ideal. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable."
CAND_0027691,19,0.940321,"NLP Engineer (Haptik, 6.5 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Vedantu.; Designed and built a significant technical initiative at Meta.; Led development of a end-to-end production system at Haptik. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions."
CAND_0081686,20,0.935280,"Search Engineer at Netflix (6.0 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Niramai.; Designed and built a significant technical initiative at Netflix.; Led development of a production system at Zoho. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions."
CAND_0044855,21,0.934269,"Senior Data Scientist (Flipkart, 6.6 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Verloop.io.; Led development of a production system at LinkedIn; Led development of an A/B testing evaluation at LinkedIn. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0057563,22,0.932876,"NLP Engineer with 6.8 years at Locobuzz, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Zoho.; Designed and built a significant technical initiative at Locobuzz.; Led development of a production system at Yellow.ai. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: could strengthen the retrieval infrastructure depth the role demands."
CAND_0018499,23,0.929386,"7.2-year Senior Machine Learning Engineer (Zomato) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Zomato.; Led development of a significant technical initiative at Flipkart.; Led development of a production system at Flipkart. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and demonstrates the founding-team ownership level the role demands. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions."
CAND_0008425,24,0.924362,"Senior NLP Engineer at Ola (7.8 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Designed and built a hybrid text search system leveraging bm25, semantic search (BM25, semantic search) at Ola.; Designed and built a significant technical initiative at Ola.; Designed and built a production system at Ola. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0088025,25,0.922136,"Staff Machine Learning Engineer (Yellow.ai, 8.6 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Designed and built a dense vector retrieval system leveraging pinecone (Pinecone) at Yellow.ai.; Designed and built a significant technical initiative at Yellow.ai.; Designed and built a end-to-end production system at Yellow.ai. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and demonstrates the founding-team ownership level the role demands. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable."
CAND_0079387,26,0.921756,"6.9-year AI Engineer at Microsoft, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at BYJU'S.; Led development of a significant technical initiative at Ola.; Led development of a end-to-end production system at Ola. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile. Consideration: career descriptions could be more explicit about the JD's must-have skills."
CAND_0077337,27,0.909353,"7.0-year Staff Machine Learning Engineer (Paytm) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Designed and built a hybrid text search system leveraging bm25, semantic search (BM25, semantic search) at Razorpay.; Designed and built a significant technical initiative at Razorpay.; Designed and built a production system at Razorpay. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: limited visible evidence on the evaluation framework design the JD emphasises."
CAND_0039838,28,0.906525,"4.9-year Recommendation Systems Engineer at Amazon, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Designed and built a significant technical initiative at Microsoft.; Designed and built a production system at Microsoft; Designed and built an A/B testing evaluation at Microsoft. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0011687,29,0.904804,7.8-year Senior NLP Engineer (Niramai) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Designed and built a dense vector retrieval system leveraging pinecone (Pinecone) at Niramai.; Designed and built a significant technical initiative at Niramai.; Designed and built a end-to-end production system at Niramai. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions.
CAND_0086022,30,0.902407,"Senior Applied Scientist at Sarvam AI (5.3 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Sarvam AI.; Designed and built a significant technical initiative at Uber.; Designed and built a production system at Uber. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: limited visible evidence on the evaluation framework design the JD emphasises."
CAND_0093912,31,0.900001,"Senior Data Scientist at Razorpay (5.3 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Flipkart.; Designed and built a significant technical initiative at Razorpay.; Led development of an evaluation framework at Flipkart. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0028793,32,0.898639,"7.2-year Search Engineer (Google) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Google.; Designed and built a significant technical initiative at Meesho.; Designed and built a production system at Meesho. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and has the architected/owned-system track record a founding role requires. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0026532,33,0.892614,"Recommendation Systems Engineer (Zomato, 4.8 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Zomato.; Led development of a production system at Mad Street Den; Led development of an A/B testing evaluation at Mad Street Den. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and brings classical IR depth AND system-ownership evidence — the JD's stated ideal. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0007412,34,0.887751,"7.4-year Applied ML Engineer at Zoho, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Zoho.; Designed and built a significant technical initiative at LinkedIn.; Designed and built a production system at Locobuzz. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0007009,35,0.882669,"7.9-year Recommendation Systems Engineer (Wysa) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Netflix.; Led development of an evaluation framework at Netflix; Led development of pre-2022 search/ranking infrastructure (elasticsearch, bm25) at Netflix. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants. Consideration: scale of prior systems is unclear from the available profile."
CAND_0075249,36,0.882551,"6.2-year Applied ML Engineer at Zomato, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at upGrad.; Led development of a production system at Zomato; Led development of an A/B testing evaluation at Zomato. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants. Consideration: evidence of A/B testing and offline-online evaluation could be stronger."
CAND_0033861,37,0.878820,"8.0-year Senior NLP Engineer (Mad Street Den) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Designed and built a hybrid text search system leveraging bm25, semantic search (BM25, semantic search) at Sarvam AI.; Designed and built a significant technical initiative at Sarvam AI.; Designed and built a production system at Sarvam AI. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: last active 121 days ago, which may indicate reduced job-search intensity."
CAND_0002025,38,0.854128,Senior AI Engineer at Apple (5.9 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a search infrastructure system leveraging re-ranking (re-ranking) at Apple.; Led development of a significant technical initiative at Apple.; Led development of a production system at Apple. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: career descriptions could be more explicit about the JD's must-have skills.
CAND_0055905,39,0.835601,"8.1-year Senior Machine Learning Engineer (Flipkart) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Rephrase.ai.; Designed and built a significant technical initiative at Flipkart.; Designed and built a production system at Flipkart. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0086151,40,0.827861,"7.7-year Recommendation Systems Engineer (Wysa) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Meta.; Designed and built a significant technical initiative at Wysa.; Led development of a production system at Glance. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: notice period of 120 days is above the JD's sub-30-day preference."
CAND_0062247,41,0.823145,"AI Engineer at Google (7.3 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Google.; Designed and built a significant technical initiative at Dream11.; Led development of an evaluation framework at Google. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: career descriptions could be more explicit about the JD's must-have skills."
CAND_0041669,42,0.809243,"Recommendation Systems Engineer (CRED, 8.0 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Mad Street Den.; Designed and built a significant technical initiative at CRED.; Led development of a production system at Yellow.ai. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: the profile would benefit from more quantified outcome evidence."
CAND_0096172,43,0.801626,"NLP Engineer (Krutrim, 5.2 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Krutrim.; Led development of a production system at Dream11; Led development of an A/B testing evaluation at Dream11. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile. Consideration: limited visible evidence on the evaluation framework design the JD emphasises."
CAND_0081053,44,0.798366,"NLP Engineer with 5.4 years at Glance, demonstrating architected/owned system leadership. Key evidence: Led development of a significant technical initiative at Glance.; Led development of a end-to-end production system at Glance; Led development of an A/B testing evaluation at Salesforce. This is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants and brings the shipped-to-real-users track record the JD's primary need demands. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0027801,45,0.798246,"NLP Engineer (InMobi, 7.4 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Zoho.; Led development of a significant technical initiative at InMobi.; Led development of a end-to-end production system at InMobi. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and brings classical IR depth AND system-ownership evidence — the JD's stated ideal. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0042029,46,0.797093,"6.5-year Senior Data Scientist at Flipkart, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Observe.AI.; Designed and built a significant technical initiative at Flipkart.; Designed and built a production system at Flipkart. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: limited visible evidence on the evaluation framework design the JD emphasises."
CAND_0037944,47,0.788525,Senior Data Scientist at Vedantu (4.9 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at PolicyBazaar.; Designed and built a significant technical initiative at Vedantu.; Designed and built a production system at Vedantu. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: the profile would benefit from more quantified outcome evidence.
CAND_0053591,48,0.786813,"5.3-year AI Engineer at Ola, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Swiggy.; Designed and built a significant technical initiative at Ola.; Designed and built a production system at Ola. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: scale of prior systems is unclear from the available profile."
CAND_0013613,49,0.777565,"4.7-year Machine Learning Engineer at Adobe, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Adobe.; Designed and built a significant technical initiative at Razorpay.; Designed and built a production system at Razorpay. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and has the architected/owned-system track record a founding role requires. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0061265,50,0.772069,"Recommendation Systems Engineer at Zoho (6.6 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Zoho.; Designed and built a significant technical initiative at Paytm.; Led development of a end-to-end production system at Observe.AI. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and demonstrates the founding-team ownership level the role demands. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0052335,51,0.771952,6.4-year Applied ML Engineer (Aganitha) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Wysa.; Designed and built a significant technical initiative at Aganitha.; Built and shipped an evaluation framework at Wysa. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: the profile would benefit from more quantified outcome evidence.
CAND_0005260,52,0.770358,"Senior NLP Engineer with 5.2 years at Netflix, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Yellow.ai.; Led development of a significant technical initiative at Netflix.; Led development of a production system at Netflix. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0077285,53,0.768209,"Recommendation Systems Engineer with 5.5 years at Nykaa, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at upGrad.; Designed and built a significant technical initiative at Nykaa.; Designed and built a production system at Nykaa. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the architected/owned-system track record a founding role requires. Consideration: career descriptions could be more explicit about the JD's must-have skills."
CAND_0051630,54,0.759239,"Machine Learning Engineer at Razorpay (6.0 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at InMobi.; Designed and built a significant technical initiative at Razorpay.; Led development of an evaluation framework at InMobi. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable."
CAND_0064270,55,0.753256,4.2-year Applied ML Engineer (Verloop.io) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Designed and built a significant technical initiative at Uber.; Designed and built a production system at Verloop.io; Designed and built an A/B testing evaluation at Verloop.io. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: could strengthen the retrieval infrastructure depth the role demands.
CAND_0022812,56,0.750762,4.5-year Applied ML Engineer (Dream11) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Designed and built a significant technical initiative at Razorpay.; Designed and built a production system at Dream11; Designed and built an A/B testing evaluation at Dream11. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions.
CAND_0054123,57,0.750665,"4.7-year Applied ML Engineer (Meta) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at CRED.; Led development of a production system at Meta; Led development of an A/B testing evaluation at Meta. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and has the evaluation rigour the JD calls out explicitly. Consideration: could strengthen the retrieval infrastructure depth the role demands."
CAND_0009691,58,0.739703,"6.2-year Applied ML Engineer at LinkedIn, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a production system at LinkedIn; Led development of an A/B testing evaluation at LinkedIn; Led development of pre-2022 search/ranking infrastructure (recommendation system, collaborative filtering) at Genpact AI. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and brings classical IR depth AND system-ownership evidence — the JD's stated ideal. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0076904,59,0.725173,"Senior Data Scientist with 4.2 years at Nykaa, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Mad Street Den.; Designed and built a significant technical initiative at Nykaa.; Led development of an evaluation framework at Mad Street Den. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and has the architected/owned-system track record a founding role requires. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0089552,60,0.722466,6.0-year Machine Learning Engineer (Netflix) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Netflix.; Led development of a production system at Mad Street Den; Led development of an A/B testing evaluation at Mad Street Den. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and brings classical IR depth AND system-ownership evidence — the JD's stated ideal. Consideration: notice period of 120 days means a slower start than the JD prefers.
CAND_0058575,61,0.721245,"5.8-year AI Engineer at Krutrim, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Krutrim.; Led development of a significant technical initiative at Uber.; Led development of a end-to-end production system at Uber. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and has the evaluation rigour the JD calls out explicitly. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0054394,62,0.710312,"4.1-year Recommendation Systems Engineer at PharmEasy, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at PharmEasy.; Designed and built a significant technical initiative at Netflix.; Led development of an evaluation framework at PharmEasy. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions."
CAND_0030827,63,0.696409,"Senior Data Scientist with 5.4 years at Freshworks, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Glance.; Designed and built a significant technical initiative at Freshworks.; Led development of an evaluation framework at Glance. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the architected/owned-system track record a founding role requires. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0070485,64,0.688042,"Search Engineer at Saarthi.ai (6.4 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Saarthi.ai.; Designed and built a significant technical initiative at Dream11.; Led development of an evaluation framework at Saarthi.ai. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and has the architected/owned-system track record a founding role requires. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0016163,65,0.684078,"Applied ML Engineer with 6.7 years at Dream11, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Verloop.io.; Led development of a significant technical initiative at Dream11.; Led development of a end-to-end production system at Dream11. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants. Consideration: notice period of 120 days is above the JD's sub-30-day preference."
CAND_0043228,66,0.681491,"Applied ML Engineer at Zoho, 6.8 years specialising in retrieval and ranking infrastructure. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Yellow.ai.; Designed and built a significant technical initiative at Zoho.; Led development of a end-to-end production system at Vedantu. This is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants and brings the shipped-to-real-users track record the JD's primary need demands. Consideration: evidence of A/B testing and offline-online evaluation could be stronger."
CAND_0076831,67,0.668145,"Search Engineer with 4.0 years at Krutrim, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Designed and built a significant technical initiative at Zomato.; Designed and built a production system at Zomato; Designed and built an A/B testing evaluation at Zomato. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the architected/owned-system track record a founding role requires. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable."
CAND_0036863,68,0.667170,"4.3-year Senior Data Scientist at upGrad, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Wysa.; Designed and built a significant technical initiative at upGrad.; Built and shipped an evaluation framework at Wysa. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0041610,69,0.665220,"Recommendation Systems Engineer (Zoho, 6.7 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at InMobi.; Designed and built a significant technical initiative at Zoho.; Led development of a end-to-end production system at Observe.AI. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and has the architected/owned-system track record a founding role requires. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0057701,70,0.657827,"Recommendation Systems Engineer (Verloop.io, 4.1 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Designed and built a significant technical initiative at Verloop.io.; Designed and built a production system at Verloop.io; Designed and built an A/B testing evaluation at Verloop.io. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: 120-day notice period exceeds the JD's preferred sub-30-day window."
CAND_0074225,71,0.652365,Machine Learning Engineer at Unacademy (4.3 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Mad Street Den.; Designed and built a significant technical initiative at Unacademy.; Designed and built a production system at Unacademy. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: notice period of 120 days is above the JD's sub-30-day preference.
CAND_0029367,72,0.652358,"Retrieval-focused Senior Data Scientist with 5.7 years of hands-on search engineering at Rephrase.ai. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Rephrase.ai.; Led development of an evaluation framework at Rephrase.ai; Led development of pre-2022 search/ranking infrastructure (elasticsearch, bm25) at CRED. This offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile and brings evaluation-framework experience the JD lists as a hard requirement. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable."
CAND_0084819,73,0.652288,"4.5-year Search Engineer at Dream11, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Dream11.; Designed and built a significant technical initiative at Razorpay.; Built and shipped an evaluation framework at Dream11. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: notice period of 120 days means a slower start than the JD prefers."
CAND_0094056,74,0.651483,"NLP Engineer at Rephrase.ai (5.9 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Rephrase.ai.; Led development of a significant technical initiative at Adobe.; Led development of a end-to-end production system at Adobe. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants. Consideration: notice period of 120 days is above the JD's sub-30-day preference."
CAND_0060054,75,0.645707,"6.4-year AI Engineer (Mad Street Den) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Mad Street Den.; Designed and built a significant technical initiative at Uber.; Designed and built a production system at Uber. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and demonstrates the founding-team ownership level the role demands. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0041568,76,0.635850,Search Engineer at Haptik (5.2 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Salesforce.; Designed and built a significant technical initiative at Haptik.; Led development of a production system at Flipkart. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable.
CAND_0083879,77,0.634923,"7.1-year Machine Learning Engineer at Ola, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Designed and built a significant technical initiative at PhonePe.; Designed and built a production system at PhonePe; Designed and built an A/B testing evaluation at PhonePe. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0052682,78,0.622449,6.6-year NLP Engineer (Aganitha) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Salesforce.; Led development of a production system at Aganitha; Led development of an A/B testing evaluation at Aganitha. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and has the evaluation rigour the JD calls out explicitly. Consideration: the profile would benefit from more quantified outcome evidence.
CAND_0010685,79,0.621389,"NLP Engineer with 6.7 years at Rephrase.ai, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Mad Street Den.; Designed and built a significant technical initiative at Rephrase.ai.; Designed and built a production system at Mad Street Den. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and has the architected/owned-system track record a founding role requires. Consideration: last active 93 days ago, which may indicate reduced job-search intensity."
CAND_0046525,80,0.608902,"Senior Machine Learning Engineer at Genpact AI (6.1 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at LinkedIn.; Designed and built a significant technical initiative at Genpact AI.; Designed and built a production system at Genpact AI. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: currently at a services/consulting firm rather than a product company."
CAND_0051292,81,0.606476,"Applied ML Engineer (Freshworks, 5.2 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Vedantu.; Designed and built a significant technical initiative at Freshworks.; Led development of an evaluation framework at Vedantu. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: career descriptions could be more explicit about the JD's must-have skills."
CAND_0020877,82,0.604737,"5.1-year Applied ML Engineer at CRED, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at CRED.; Led development of an evaluation framework at CRED; Spent 5.0 of 5.0 total career years in the JD's core domain. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and brings evaluation-framework experience the JD lists as a hard requirement. Consideration: career descriptions could be more explicit about the JD's must-have skills."
CAND_0036184,83,0.594239,"Search-and-ranking Recommendation Systems Engineer (CRED), 6.0 years of retrieval-system engineering. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at CRED.; Led development of an evaluation framework at CRED; Led development of pre-2022 search/ranking infrastructure (elasticsearch, bm25) at PolicyBazaar. This brings classical IR depth AND system-ownership evidence — the JD's stated ideal. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions."
CAND_0098846,84,0.592046,7.6-year AI Engineer whose career centres on search and ranking systems. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Swiggy.; Designed and built a significant technical initiative at upGrad.; Led development of a end-to-end production system at Meesho. This has the architected/owned-system track record a founding role requires and brings evaluation-framework experience the JD lists as a hard requirement. Consideration: depth in vector-database operations is not clearly evidenced in career descriptions.
CAND_0094759,85,0.580607,"8.6-year Lead AI Engineer (Meta) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Led development of a hybrid vector search system leveraging bm25, faiss, dense retrieval, +2 more (BM25, FAISS, dense retrieval) at Apple.; Designed and built a significant technical initiative at Meta.; Designed and built a production system at Meta. This matches the JD's Tier-5 ideal — strong product-company career evidence even without leaning on JD-specific keywords and demonstrates the founding-team ownership level the role demands. Consideration: profile inactive for 160 days combined with weak behavioural signals — the JD's 'not actually available' trap."
CAND_0030348,86,0.564244,"Machine Learning Engineer at BYJU'S, 4.5 years specialising in retrieval and ranking infrastructure. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at BYJU'S.; Led development of a production system at PolicyBazaar; Led development of an A/B testing evaluation at PolicyBazaar. This brings evaluation-framework experience the JD lists as a hard requirement and has the production shipping record essential for this role. Consideration: could strengthen the retrieval infrastructure depth the role demands."
CAND_0049896,87,0.563577,"Search Engineer (Unacademy, 7.3 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Unacademy.; Led development of a production system at Flipkart; Led development of an A/B testing evaluation at Flipkart. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and is the rare pre-2022 IR + senior-ownership combination the JD explicitly wants. Consideration: notice period of 90 days may slow the start the JD wants."
CAND_0079064,88,0.563017,"5.2-year Senior Data Scientist at Niramai, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Niramai.; Led development of an evaluation framework at Niramai; Spent 5.2 of 5.2 total career years in the JD's core domain. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows NDCG/MRR/MAP experience the JD ranks as a must-have. Consideration: notice period of 120 days is above the JD's sub-30-day preference."
CAND_0050454,89,0.561888,"6.8-year AI Engineer at Rephrase.ai, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Rephrase.ai.; Designed and built a significant technical initiative at Uber.; Led development of an evaluation framework at Rephrase.ai. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and shows the end-to-end ownership depth expected of a founding-team hire. Consideration: production-system scale evidence is thinner than the JD's 'meaningful scale' bar."
CAND_0066376,90,0.553387,"Applied ML Engineer (Dream11, 5.7 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Salesforce.; Led development of a production system at Dream11; Led development of an A/B testing evaluation at Dream11. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile. Consideration: 90-day notice period is above the JD's preferred sub-30-day window but workable."
CAND_0031593,91,0.549923,"Search Engineer with 7.8 years at Genpact AI, a Tier-5 profile whose career history speaks louder than their skill list. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Unacademy.; Designed and built a significant technical initiative at Genpact AI.; Led development of a production system at Aganitha. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the architected/owned-system track record a founding role requires. Consideration: career spans many unrelated domains, raising depth concerns."
CAND_0020708,92,0.541217,"4.2-year Search Engineer whose career centres on search and ranking systems. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at PolicyBazaar.; Led development of an evaluation framework at PolicyBazaar; Spent 4.2 of 4.2 total career years in the JD's core domain. This partially aligns with the JD's core technical requirements. Consideration: could strengthen the retrieval infrastructure depth the role demands."
CAND_0003977,93,0.538628,"Recommendation Systems Engineer (Google) with 4.6 years building retrieval infrastructure at scale. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Google.; Led development of an evaluation framework at Google; Led development of pre-2022 search/ranking infrastructure (elasticsearch, bm25) at Rephrase.ai. This partially aligns with the JD's core technical requirements. Consideration: last active 91 days ago, which may indicate reduced job-search intensity."
CAND_0040117,94,0.536430,"Recommendation Systems Engineer (PhonePe, 6.5 yrs) whose career depth fits the JD's Tier-5 ideal even without perfect keyword coverage. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Salesforce.; Led development of a production system at PhonePe; Led development of an A/B testing evaluation at PhonePe. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the evaluation rigour the JD calls out explicitly. Consideration: last active 92 days ago, which may indicate reduced job-search intensity."
CAND_0040887,95,0.536426,"Search-and-ranking Machine Learning Engineer (Netflix), 4.7 years of retrieval-system engineering. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Netflix.; Led development of an evaluation framework at Netflix; Spent 4.6 of 4.6 total career years in the JD's core domain. This brings evaluation-framework experience the JD lists as a hard requirement. Consideration: the profile would benefit from more quantified outcome evidence."
CAND_0070398,96,0.536415,"7.2-year Machine Learning Engineer at Genpact AI, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Led development of a hybrid vector search system leveraging bm25, elasticsearch, faiss, +1 more (BM25, Elasticsearch, FAISS) at Genpact AI.; Designed and built a significant technical initiative at Mad Street Den.; Designed and built a production system at Mad Street Den. This exemplifies the Tier-5 profile the JD describes: career history outranks keyword coverage and offers pre-LLM IR grounding combined with senior ownership — exactly the JD's 'understood retrieval before it became fashionable' profile. Consideration: notice period of 120 days is above the JD's sub-30-day preference."
CAND_0075439,97,0.524830,Machine Learning Engineer at Flipkart (4.3 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Flipkart.; Led development of a production system at Genpact AI; Led development of an A/B testing evaluation at Genpact AI. This is the kind of Tier-5 candidate the JD explicitly wants — ownership and shipped systems over keyword lists and has the evaluation rigour the JD calls out explicitly. Consideration: LangChain-heavy recent work without production deployment evidence — the JD's 'framework enthusiasts' trap.
CAND_0032807,98,0.523213,"4.2-year Machine Learning Engineer at CRED, the kind of 'built-it-at-a-product-company' profile the JD explicitly calls out. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at CRED.; Built and shipped an evaluation framework at CRED; Spent 4.2 of 4.2 total career years in the JD's core domain. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: LangChain-heavy recent work without production deployment evidence — the JD's 'framework enthusiasts' trap."
CAND_0042506,99,0.523148,Search Engineer at Verloop.io (4.2 yrs) with a Tier-5 product-engineer signature: shipped and owned systems without leaning on JD keywords. Key evidence: Built and shipped a dense vector retrieval system leveraging pinecone (Pinecone) at Verloop.io.; Led development of a production system at Meesho; Led development of an A/B testing evaluation at Meesho. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and brings evaluation-framework experience the JD lists as a hard requirement. Consideration: LangChain-heavy recent work without production deployment evidence — the JD's 'framework enthusiasts' trap.
CAND_0007411,100,0.523146,8.0-year Senior Machine Learning Engineer (Amazon) whose career evidence outweighs the absence of explicit JD keywords. Key evidence: Designed and built a dense vector retrieval system leveraging pinecone (Pinecone) at Amazon.; Designed and built a significant technical initiative at Amazon.; Designed and built a end-to-end production system at Amazon. This fits the JD's 'built-it-at-a-product-company' Tier-5 profile that the spec explicitly calls out and demonstrates the founding-team ownership level the role demands. Consideration: profile inactive for 197 days combined with weak behavioural signals — the JD's 'not actually available' trap.