Skip to content

RabbaniHacker/shadowintern

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🕵️ ShadowIntern

Train for ONE specific company's exact culture and problems — before you ever apply.

Status License Built With


🧠 The Core Insight

Every top company has a unique engineering culture — how they write code, what problems they solve, how they review work, what they value in engineers.

The candidate who walks in already thinking like a Razorpay engineer always beats the one who just grinded DSA.

ShadowIntern makes you that person — for any company, without knowing anyone inside.


🚀 How It Works

Student enters: "I want to work at Razorpay"
        │
        ▼
┌─────────────────────────────────────┐
│  Agent 1 — Company Intelligence     │
│  Scrapes GitHub repos, tech blogs,  │
│  job descriptions, Glassdoor,       │
│  LinkedIn posts, YouTube talks      │
│  → Returns: "Company DNA"           │
└──────────────┬──────────────────────┘
               │
        ▼
┌─────────────────────────────────────┐
│  Agent 2 — Culture Modeller         │
│  Understands stack, values,         │
│  what they reject, how they think   │
│  → Returns: "Engineer Profile"      │
└──────────────┬──────────────────────┘
               │
        ▼
┌─────────────────────────────────────┐
│  Agent 3 — Task Generator           │
│  Creates REAL intern-level tasks    │
│  based on actual company problems   │
│  → Returns: Daily task assignments  │
└──────────────┬──────────────────────┘
               │
        ▼
┌─────────────────────────────────────┐
│  Agent 4 — Shadow Mentor            │
│  Reviews code like a senior         │
│  engineer AT THAT COMPANY would     │
│  → Returns: Company-style feedback  │
└──────────────┬──────────────────────┘
               │
        ▼
┌─────────────────────────────────────┐
│  Agent 5 — Interview Simulator      │
│  Uses ACTUAL questions their        │
│  interviewers ask, gives verdict    │
│  → Returns: Hire / No-hire + gaps   │
└─────────────────────────────────────┘

📊 Dashboard Preview (Concept)

🏢 Shadow Company: Razorpay          📅 Week 3 of 4
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

✅ Tasks Completed:     11/14
⭐ Senior Review Score: 74/100
🎯 Interview Readiness: 68%

📋 This Week's Task:
   "Design a webhook retry system that handles
    partial failures gracefully.
    Razorpay processes 5M webhooks/day."

👨‍💼 Mentor Feedback:
   "Your solution works but a Razorpay engineer
    would flag this — you're not handling
    idempotency. That's a P0 bug in payments.
    Redo section 3."

📈 Weak Area:   System Design — Distributed Queues
🔥 Strong Area: API Design, Error Handling

🆚 Why This Is Different

What Exists What ShadowIntern Does
LeetCode Generic DSA, no company context
Mock interview platforms Scripted questions, not company-specific
Internships Luck-based, requires connections
ShadowIntern Train for ONE company's exact culture & problems

🛠️ Tech Stack

Layer Tool
Agent Framework CrewAI + LangGraph
LLM Claude API (Anthropic)
Scraping Playwright + GitHub API
Code Review Agent Claude with company style prompts
Frontend React Dashboard
Auth + Payments Supabase + Razorpay (ironic)

📈 Scale Roadmap

See ROADMAP.md for full details.

  • Phase 1 → Top 20 Indian tech companies
  • Phase 2 → Any company globally with public data
  • Phase 3 → ₹500/month per company (student plan)
  • Phase 4 → Companies pay to access pre-trained candidates

🔬 Research

This project also has an active research component exploring multi-agent frameworks for company-specific engineering culture extraction and task synthesis.

See research/ABSTRACT.md for the working paper draft.

Related work:

  • MockLLM (2024) — Multi-agent mock interview generation
  • TheAgentCompany (2024) — Benchmarking LLM agents on real workplace tasks
  • MetaGPT (2023) — Multi-agent software development simulation

🗺️ Architecture

See ARCHITECTURE.md for the full agent pipeline breakdown.


🤝 Contributing

This project is in early design phase. See CONTRIBUTING.md if you want to collaborate.

Areas where help is needed:

  • Agent pipeline design
  • Company data scraping strategy
  • Frontend dashboard mockups
  • Research/evaluation methodology

👤 Author

Built and researched by @RabbaniHacker


📄 License

MIT © 2026 — See LICENSE

About

AI that gets you hired by making you work at a company before you apply

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors