class RudraDudhat:
institute = "Indian Institute of Technology, Bhilai"
degree = "B.Tech — Data Science & Artificial Intelligence"
cgpa = 9.48 # out of 10
year = "2nd Year (2024–2028)"
location = "Navi Mumbai, India"
shipped = {
"Cascade AI" : "Firebase Cloud Functions — Multi-Agent LLM Pipeline",
"OptiQuant" : "Streamlit Cloud — Full-Stack ML Trading App",
}
stack = {
"LLM" : ["LangChain", "OpenAI API", "HuggingFace", "RAG", "Ollama"],
"ML" : ["LightGBM", "CatBoost", "PyTorch", "SHAP", "Scikit-learn"],
"Infra" : ["Docker", "AWS EC2", "Firebase", "GitHub Actions", "CI/CD"],
"Web" : ["Streamlit", "Flask", "Gradio"],
}
philosophy = "A model that runs only on your laptop is not a product."
seeking = "AI Engineering Internship — Remote / Mumbai"
def ship(self, idea):
prototype = self.build(idea)
containerized = self.dockerize(prototype)
deployed = self.push_to_cloud(containerized)
return deployed # not prototype. deployed.
An LLM planner that decomposes plain-text goals into coordinated multi-agent blueprints — web search, scraping, auditing, routing, and formatting — executed in sequence via Firebase Cloud Functions. Self-healing router with conditional branching handles mid-run failures automatically. Per-user private memory stored in Firestore; large intermediate payloads offloaded to GCS.
Not a chatbot. Not an API wrapper. A production orchestration system.
Weighted ensemble (LightGBM 40% + CatBoost 40% + Random Forest 20%) with walk-forward backtesting, SHAP explainability surfaced directly in the UI, and full Docker + GitHub Actions CI/CD. Found lookahead bias in v1 — rebuilt from scratch instead of hiding it.
| Metric | Result |
|---|---|
| Sharpe Ratio | 1.05 |
| CAGR | 28.16% |
| Win Rate | 53.92% |
| Max Drawdown | -20.00% |
| Calmar Ratio | 1.41 |
Scrapes job listings, scores them using embedding similarity against your profile, generates personalized cold emails with GPT-4o, requires human approval before sending, then classifies replies and retrains a logistic regression model to re-rank future opportunities. Gets smarter with every send.
LANGUAGES
ML / DEEP LEARNING
LLM & GENAI
INFRA & DEVOPS
Powered by WakaTime — updates every 24 hours via GitHub Actions
🐱 My GitHub Data
📦 ? Used in GitHub's Storage
🏆 116 Contributions in the Year 2026
💼 Opted to Hire
📜 18 Public Repositories
🔑 0 Private Repositories
I'm an Early 🐤
🌞 Morning 10 commits █░░░░░░░░░░░░░░░░░░░░░░░░ 05.00 %
🌆 Daytime 95 commits ████████████░░░░░░░░░░░░░ 47.50 %
🌃 Evening 68 commits ████████░░░░░░░░░░░░░░░░░ 34.00 %
🌙 Night 27 commits ███░░░░░░░░░░░░░░░░░░░░░░ 13.50 %
📅 I'm Most Productive on Saturday
Monday 5 commits █░░░░░░░░░░░░░░░░░░░░░░░░ 02.50 %
Tuesday 53 commits ███████░░░░░░░░░░░░░░░░░░ 26.50 %
Wednesday 16 commits ██░░░░░░░░░░░░░░░░░░░░░░░ 08.00 %
Thursday 33 commits ████░░░░░░░░░░░░░░░░░░░░░ 16.50 %
Friday 8 commits █░░░░░░░░░░░░░░░░░░░░░░░░ 04.00 %
Saturday 65 commits ████████░░░░░░░░░░░░░░░░░ 32.50 %
Sunday 20 commits ██░░░░░░░░░░░░░░░░░░░░░░░ 10.00 %
📊 This Week I Spent My Time On
💬 Programming Languages:
Python 8 mins ██████████████████░░░░░░░ 70.97 %
HTML 3 mins ███████░░░░░░░░░░░░░░░░░░ 29.03 %
I Mostly Code in Jupyter Notebook
Jupyter Notebook 8 repos ████████████░░░░░░░░░░░░░ 50.00 %
Python 3 repos █████░░░░░░░░░░░░░░░░░░░░ 18.75 %
HTML 3 repos █████░░░░░░░░░░░░░░░░░░░░ 18.75 %
TypeScript 2 repos ███░░░░░░░░░░░░░░░░░░░░░░ 12.50 %
Timeline
Last Updated on 24/03/2026 01:48:12 UTC
Aug 2025 – Present Core Member DS&AI Club, IIT Bhilai
Led 12-person team 200+ students across workshops
Aug 2025 – Present Placement Volunteer CCPS, IIT Bhilai
50+ cold outreach 2 new recruiters onboarded
Jul 2024 – May 2028 B.Tech DSAI IIT Bhilai
CGPA 9.48 / 10 JEE Advanced Rank 10,463
"A model that runs only on your laptop is not a product. It is a proof of concept wearing a costume."