
Hi, I'mPratham Patil
Pune, India
AI & ML engineering graduate specializing in Generative AI — building end-to-end LLM systems across RAG pipelines, agentic document workflows, and production observability stacks. Proficient with the modern GenAI toolchain: LangChain, LangGraph, ChromaDB, FastAPI and GitHub Actions CI/CD.
Building AI that ships to production
B.E. – Artificial Intelligence & Machine Learning
PES's Modern College of Engineering, Pune
My Story
I'm an AI & ML engineering graduate from PES's Modern College of Engineering, Pune, specializing in Generative AI. I build end-to-end LLM systems — from RAG pipelines and agentic document workflows to production observability stacks.
My work spans the modern GenAI toolchain: LangChain, LangGraph, ChromaDB, FastAPI, and GitHub Actions CI/CD. I've shipped projects like CodeNarrator, an autonomous documentation and video synthesis system, and a full LLM observability pipeline with RAGAS-based CI/CD evaluation.
I believe in building AI products that work in production, not just in notebooks. Every project I take on is driven by real-world utility — measurable performance, clean architecture, and observable systems.
Experience
Artificial Intelligence Intern
Aroma Brand Solutions
- Performed data preprocessing, feature engineering, and EDA on marketing and user engagement datasets using Pandas
- Surfaced data-driven insights that informed branding and campaign strategies
- Collaborated with marketing, design, and web teams to apply ML-driven analytics to digital content strategy
Technical Expertise
6 categories · 49 technologies across AI, backend, cloud & data
Programming Languages
Generative AI & LLMs
RAG & Agentic Systems
Eval & Observability
APIs & Frameworks
MLOps, Cloud & Data
Featured Projects
7 projects · AI · Data · Full-Stack
Code Narrator
Autonomous Documentation & Video Synthesis System
An AI-powered system that automatically transforms GitHub repositories into comprehensive video tutorials with chapter-by-chapter markdown documentation and fully narrated MP4 videos — no manual intervention required.
RetrieveX
LLM-Powered Intelligent Query Retrieval System
A production-ready RAG system that processes large PDF documents and delivers intelligent, contextual answers for insurance, legal, HR, and compliance use cases using semantic search and GPT-4.
YouTube Q&A
RAG-Based Video Intelligence System
A RAG pipeline that transcribes YouTube audio via Whisper, indexes timestamped chunks into ChromaDB, and answers natural language queries with exact video moment citations.
AI Performance Monitor
LLM Observability & CI/CD Evaluation Pipeline
A production LLM observability system that captures 9 metrics per query — latency, token usage, cost, faithfulness, context recall, quality drift — with automated CI/CD evaluation via GitHub Actions.
Get in touch
Pratham Patil
Generative AI Engineer
Usually responds within 24 hours. Happy to discuss roles, projects, or research ideas.