B.Tech student (AI & ML) building RAG and agentic AI systems. Python · LangGraph · ChromaDB · Groq · Flask · Next.js

Open to AI / GenAI internships.

A multi-agent LangGraph pipeline that turns a topic into a narrated, animated explainer video.

• Planner → Narration → Visual → Assembler agents, run as a structured pipeline
• Narration is generated first, and its measured audio duration paces the animation
• The visual agent writes Manim code and retries with the real render error when a render fails
• Async job processing with SQLite job tracking, FFmpeg assembly, Docker setup
• Honest Known Limitations section in the repo

LangGraph Groq Manim Piper TTS FFmpeg Flask Next.js SQLite Docker

Repository · Live frontend

A backend-only study tool that covers the GenAI stack end to end: structured output, RAG, a tool-calling agent and evaluation.

• RAG over uploaded PDFs: chunking with overlap, local embeddings, ChromaDB, similarity-threshold filtering
• LangGraph ReAct agent that decides which tool to use, e.g. "quiz me on my weak topics"
• Schema-enforced quizzes via Groq Structured Outputs, plus SQLite weak-topic tracking
• LLM-as-a-judge faithfulness check to catch answers not grounded in the retrieved context

LangGraph ChromaDB Sentence-Transformers Groq Pydantic Flask SQLite

Repository · Live API (free tier, may be slow on first request)

An AI-powered Delhi Metro platform with route planning, ML-based crowd prediction, fare estimation, digital tickets, chatbot assistance and nearby attractions.

Next.js Flask Python Machine Learning

An AI companion for introverted and socially anxious students: mood-aware chat, journaling, challenges and music/movie recommendations, with a crisis-safety branch that skips the AI and shows a helpline.

Next.js Flask LangGraph Groq

• Building an agentic AI system for air-quality hazard forecasting and personalized risk explanation (repo coming soon)
• Strengthening DSA: arrays, strings, trees, graphs, dynamic programming
• Learning AI evaluation and cloud deployment