Kunjan Shah commited on
Commit
1fea7cc
·
1 Parent(s): d1a3298
Files changed (1) hide show
  1. script.md +60 -0
script.md ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Demo Script: HiredGPT
2
+
3
+ DuelHiredGPT Duel – Competitive Interview Simulation
4
+
5
+ Hello everyone,
6
+
7
+ I’m Amal, and this is HiredGPT Duel – a competitive interview simulator that shows you exactly how you stack up against a top-tier candidate.
8
+
9
+ 1 — Problem & Idea\
10
+ Most interview-prep tools grade you in isolation. You get a score, but no sense of what great really looks like.
11
+ HiredGPT Duel fixes that by pitting you against an AI “star applicant” who’s about 10 percent stronger than your own résumé. After every answer you see a side-by-side comparison and immediate feedback.
12
+
13
+ 2 — Quick Walk-through\
14
+ 1. Upload your resume and the target job description in pdf format or paste job description text format.
15
+
16
+ 2. Pick how much stronger the rival should be—10 to 100 percent—or leave it at the default 10 percent.
17
+
18
+ 3. Enter the job role you are targeting
19
+
20
+ 4. Submit to generate rival resume and interview questions.
21
+
22
+ 5. Behind the scenes, the system generates (LLM) interview questions based on submited resume JD and job description with few shot prompt engineering tech.
23
+
24
+ 6. System generates difficulty level, job desciption, resume and key factors exteacted.
25
+ Now we jump into the duel.
26
+
27
+ - Question view
28
+ on the top
29
+
30
+ - Answer the question—type or speak. If you record audio, Whisper transcribes it instantly.
31
+ - Click “Generate Rival Answer.” The AI competitor responds using the upgraded résumé.
32
+ -
33
+
34
+
35
+
36
+
37
+
38
+
39
+
40
+
41
+
42
+
43
+
44
+
45
+
46
+
47
+ 3 — Tech Highlights
48
+ Frontend: Streamlit for a fast, sharable UI.
49
+
50
+ LLMs: Google Gemini (via Groq client) for question generation, résumé enhancement, rival answers, and scoring and report generation.
51
+
52
+ Speech-to-Text: OpenAI Whisper running locally for offline transcription
53
+
54
+ Text-to-Speech: pyttsx3 python library. if you’d like the rival to read answers aloud.
55
+
56
+ Document Handling: PyPDF2 for parsing PDFs.
57
+ Everything runs on-device except the LLM calls, keeping latency low and privacy high.
58
+
59
+
60
+