hamzashahid40 commited on
Commit
1838104
·
verified ·
1 Parent(s): c53b2ba

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +104 -4
README.md CHANGED
@@ -1,10 +1,110 @@
1
  ---
2
  title: README
3
- emoji: 🐢
4
- colorFrom: pink
5
- colorTo: blue
6
  sdk: static
7
  pinned: false
8
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9
 
10
- Edit this `README.md` markdown file to author your organization card.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  title: README
3
+ emoji: 🤖
4
+ colorFrom: blue
5
+ colorTo: green
6
  sdk: static
7
  pinned: false
8
  ---
9
+ <style>
10
+ img.rotating-content {
11
+ width: 300px;
12
+ height: 150px;
13
+ object-fit:cover;
14
+ object-position:center;
15
+ }
16
+ .icon {
17
+ width: 1.5em;
18
+ height: 1.5em;
19
+ vertical-align: -.7em;
20
+ padding: .25em .5em .25em .25em;
21
+ }
22
+ ul.social {
23
+ list-style: none;
24
+ padding-left: 0;
25
+ }
26
+ ul.social li {
27
+ padding-left: .5em;
28
+ display: flex;
29
+ }
30
+ .architectureImage {
31
+ border: 0.5px solid black;
32
+ }
33
+ </style>
34
+ <img src="images/logo.png" alt="Navid Banner"/>
35
+ _**Building Arabic-first AI with Navid**_
36
 
37
+ At **Navid**, we design, train, and deploy Arabic-first and multilingual AI systems. From curated data pipelines and synthetic data generation to custom LLM training, evaluation, and production-grade RAG/agentic apps, we help teams in KSA and beyond ship robust AI—on-prem or in the cloud.
38
+
39
+ ## Build and Scale GenAI with Navid
40
+
41
+ Navid offers an end-to-end stack spanning data, models, and applications. We specialize in:
42
+
43
+ - **Data pipelines & governance** — automated collection, cleansing, deduping, safety filtering, and curation.
44
+ - **Synthetic data & augmentation** — domain-balanced recipe design for low-resource Arabic tasks.
45
+ - **Training & fine-tuning** — SFT, DPO/PPO, curriculum & multi-stage training on open and custom architectures.
46
+ - **Custom LLMs & evals** — Arabic/multilingual capabilities, eval harnesses, human+AI QA, and enterprise red-teaming.
47
+ - **Advanced RAG** — document intelligence, retriever/re-ranker selection, vector/hybrid search, caching & guardrails.
48
+ - **Agentic workflows** — LangGraph/LangChain-based multi-agent systems with tools, memory, and supervision.
49
+ - **Secure deployment** — on-prem, VPC, or managed GPU (e.g., vLLM, Triton) with observability (Prometheus/Grafana/Jaeger).
50
+
51
+ Whether you need domain-tuned assistants, multilingual search, or end-to-end AI transformation aligned with **Vision 2030**, Navid provides production-grade solutions tailored for enterprise scale and policy compliance.
52
+
53
+ Learn more about our capabilities:
54
+
55
+ * Arabic-first **LLM Training & Fine-tuning**
56
+ * **On‑prem & VPC RAG** for regulated data
57
+ * **Evaluation & Benchmarks** for Arabic/multilingual tasks
58
+ * **Agentic Applications** for workflows & decision support
59
+
60
+ > _Ask us about Navid Hub, our foundation for dataset/versioning, experiment tracking, and deployment automation._
61
+
62
+ ## Building AI Agents with Navid
63
+
64
+ We design multi-agent systems for retrieval, planning, reasoning, and actions across enterprise tools. Our approach combines:
65
+
66
+ - **Structured tool-use** with typed actions and safety constraints
67
+ - **Planner–Worker patterns** using LangGraph
68
+ - **Grounded answering** via graded hallucination control
69
+ - **Replayable traces & guardrails** for compliance and audit
70
+
71
+ Use cases: Arabic knowledge copilots, policy Q&A, financial analytics, code assistants, and L4/L5 support automation.
72
+
73
+ * Get started with **Agentic Templates** (LangGraph/LangChain)
74
+ * Production **Evaluation Harness** for agents (trace-based)
75
+
76
+ ## Navid & Open Source
77
+
78
+ We contribute to the Arabic AI ecosystem: datasets, re-rankers for RAG, evaluation recipes, and MLOps utilities. Highlights include:
79
+
80
+ - **Arabic/Multilingual Reranker Leaderboard**
81
+ - **RAG Evaluation Recipes** (faithfulness, context precision/recall, semantic similarity)
82
+ - **Data loaders & cleaners** for Arabic PDFs and OCR post-processing
83
+
84
+ * Explore sample repos and templates (fill in links below):
85
+ * GitHub: `<your_org_url_here>`
86
+ * Docs: `<your_docs_url_here>`
87
+
88
+ ## Connect, Learn, and Grow with Navid
89
+
90
+ <div class="grid lg:grid-cols-2 gap-x-4 gap-y-7 p2">
91
+ <div class="col-span-1 p2">
92
+ <p>Navid Community</p>
93
+ <ul class="social">
94
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512" class="icon"><path fill="currentColor" d="M448 96c0-35.3-28.7-64-64-64H64C28.7 32 0 60.7 0 96V416c0 35.3 28.7 64 64 64H384c35.3 0 64-28.7 64-64V96zM265.8 407.7c0-1.8 0-6 .1-11.6c.1-11.4 .1-28.8 .1-43.7c0-15.6-5.2-25.5-11.3-30.7c37-4.1 76-9.2 76-73.1c0-18.2-6.5-27.3-17.1-39c1.7-4.3 7.4-22-1.7-45c-13.9-4.3-45.7 17.9-45.7 17.9c-13.2-3.7-27.5-5.6-41.6-5.6s-28.4 1.9-41.6 5.6c0 0-31.8-22.2-45.7-17.9c-9.1 22.9-3.5 40.6-1.7 45c-10.6 11.7-15.6 20.8-15.6 39c0 63.6 37.3 69 74.3 73.1c-4.8 4.3-9.1 11.7-10.6 22.3c-9.5 4.3-33.8 11.7-48.3-13.9c-9.1-15.8-25.5-17.1-25.5-17.1c-16.2-.2-1.1 10.2-1.1 10.2c10.8 5 18.4 24.2 18.4 24.2c9.7 29.7 56.1 19.7 56.1 19.7c0 9 .1 21.7 .1 30.6c0 4.8 .1 8.6 .1 10c0 4.3-3 9.5-11.5 8C106 393.6 59.8 330.8 59.8 257.4c0-91.8 70.2-161.5 162-161.5s166.2 69.7 166.2 161.5c.1 73.4-44.7 136.3-110.7 158.3c-8.4 1.5-11.5-3.7-11.5-8z"/></svg><a href="#">Navid on GitHub</a></li>
95
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512" class="icon"><path fill="currentColor" d="M416 32H31.9C14.3 32 0 46.5 0 64.3v383.4C0 465.5 14.3 480 31.9 480H416c17.6 0 32-14.5 32-32.3V64.3c0-17.8-14.4-32.3-32-32.3zM135.4 416H69V202.2h66.5V416zm-33.2-243c-21.3 0-38.5-17.3-38.5-38.5S80.9 96 102.2 96c21.2 0 38.5 17.3 38.5 38.5 0 21.3-17.2 38.5-38.5 38.5zm282.1 243h-66.4V312c0-24.8-.5-56.7-34.5-56.7-34.6 0-39.9 27-39.9 54.9V416h-66.4V202.2h63.7v29.2h.9c8.9-16.8 30.6-34.5 62.9-34.5 67.2 0 79.7 44.3 79.7 101.9V416z"/></svg><a href="#">Navid on LinkedIn</a></li>
96
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 640 512" class="icon"><path fill="currentColor" d="M180.4 203c-.7 22.7 10.6 32.7 10.9 39.1a8.2 8.2 0 0 1 -4.1 6.3l-12.8 9a10.7 10.7 0 0 1 -5.6 1.9c-.4 0-8.2 1.8-20.5-25.6a78.6 78.6 0 0 1 -62.6 29.5c-16.3 .9-60.4-9.2-58.1-56.2-1.6-38.3 34.1-62.1 70.9-60.1 7.1 0 21.6 .4 47 6.3v-15.6c2.7-26.5-14.7-47-44.8-43.9-2.4 0-19.4-.5-45.8 10.1-7.4 3.4-8.3 2.8-10.8 2.8-7.4 0-4.4-21.5-2.9-24.2 5.2-6.4 35.9-18.4 65.9-18.2a76.9 76.9 0 0 1 55.7 17.3 70.3 70.3 0 0 1 17.7 52.4l0 69.3z"/></svg><a href="#">Open Source at Navid</a></li>
97
+ </ul>
98
+ </div>
99
+ <div class="col-span-1 p2">
100
+ <p>Research & Datasets</p>
101
+ <ul class="social">
102
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 448 512" class="icon"><path fill="currentColor" d="M448 96c0-35.3-28.7-64-64-64H64C28.7 32 0 60.7 0 96V416c0 35.3 28.7 64 64 64H384c35.3 0 64-28.7 64-64V96zM265.8 407.7c0-1.8 0-6 .1-11.6c.1-11.4 .1-28.8 .1-43.7c0-15.6-5.2-25.5-11.3-30.7c37-4.1 76-9.2 76-73.1c0-18.2-6.5-27.3-17.1-39c1.7-4.3 7.4-22-1.7-45c-13.9-4.3-45.7 17.9-45.7 17.9c-13.2-3.7-27.5-5.6-41.6-5.6s-28.4 1.9-41.6 5.6c0 0-31.8-22.2-45.7-17.9c-9.1 22.9-3.5 40.6-1.7 45c-10.6 11.7-15.6 20.8-15.6 39c0 63.6 37.3 69 74.3 73.1c-4.8 4.3-9.1 11.7-10.6 22.3c-9.5 4.3-33.8 11.7-48.3-13.9c-9.1-15.8-25.5-17.1-25.5-17.1c-16.2-.2-1.1 10.2-1.1 10.2c10.8 5 18.4 24.2 18.4 24.2c9.7 29.7 56.1 19.7 56.1 19.7c0 9 .1 21.7 .1 30.6c0 4.8 .1 8.6 .1 10c0 4.3-3 9.5-11.5 8C106 393.6 59.8 330.8 59.8 257.4c0-91.8 70.2-161.5 162-161.5s166.2 69.7 166.2 161.5c.1 73.4-44.7 136.3-110.7 158.3c-8.4 1.5-11.5-3.7-11.5-8z"/></svg><a href="#">Navid Research on GitHub</a></li>
103
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon"><path fill="currentColor" d="M64 32C28.7 32 0 60.7 0 96L0 416c0 35.3 28.7 64 64 64l384 0c35.3 0 64-28.7 64-64l0-320c0-35.3-28.7-64-64-64L64 32zm88 64l0 64-88 0 0-64 88 0zm56 0l88 0 0 64-88 0 0-64zm240 0l0 64-88 0 0-64 88 0zM64 224l88 0 0 64-88 0 0-64zm232 0l0 64-88 0 0-64 88 0zm64 0l88 0 0 64-88 0 0-64zM152 352l0 64-88 0 0-64 88 0zm56 0l88 0 0 64-88 0 0-64zm240 0l0 64-88 0 0-64 88 0z"/></svg><a href="#">Code & Datasets</a></li>
104
+ </ul>
105
+ </div>
106
+ <div class="col-span-1 p2">
107
+ <p>Navid Developers</p>
108
+ <ul class="social">
109
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon"><path fill="currentColor" d="M192 32c0 17.7 14.3 32 32 32c123.7 0 224 100.3 224 224c0 17.7 14.3 32 32 32s32-14.3 32-32C512 128.9 383.1 0 224 0c-17.7 0-32 14.3-32 32zm0 96c0 17.7 14.3 32 32 32c70.7 0 128 57.3 128 128c0 17.7 14.3 32 32 32s32-14.3 32-32c0-106-86-192-192-192c-17.7 0-32 14.3-32 32zM96 144c0-26.5-21.5-48-48-48S0 117.5 0 144L0 368c0 79.5 64.5 144 144 144s144-64.5 144-144s-64.5-144-144-144l-16 0 0 96 16 0c26.5 0 48 21.5 48 48s-21.5 48-48 48s-48-21.5-48-48l0-224z"/></svg><a href="#">Navid Engineering Blog</a></li>
110
+ <li><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 512 512" class="icon"><path fill="currentColor" d="M184 0c30.9 0 56 25.1 56 56l0