I design and build AI-powered applications, agentic systems, and RAG pipelines with a focus on turning LLM capabilities into reliable, deployable software.
My work spans agent orchestration, retrieval pipelines, backend systems, APIs, and full-stack applications — from experimentation and evaluation to deployment.
- Agentic AI Systems — multi-step workflows, tool-using agents, autonomous task execution, and stateful AI applications
- RAG Systems — document ingestion, preprocessing, embeddings, retrieval, context construction, and grounded generation
- AI Applications — LLM-powered products built around practical user workflows rather than simple chat interfaces
- Backend Systems — Python, FastAPI, databases, APIs, authentication, and service architecture
- Full-Stack Applications — Next.js, TypeScript, and modern web applications for deploying AI systems
AI / ML: LangGraph, LangChain, RAG, LLMs, NLP, TensorFlow Backend: Python, FastAPI, REST APIs, PostgreSQL, MongoDB Frontend: Next.js, TypeScript, React, Tailwind CSS Tools & Infrastructure: Git, GitHub, Docker, Vercel, Render
I learn by building real systems, debugging what breaks, and shipping what works.
-->
AI-powered market intelligence and saturation analysis system.
An agentic workflow that evaluates product ideas by researching existing products, analyzing the competitive landscape, and generating a structured saturation assessment.
Built with: Python, LangGraph, LLMs, web search, agentic workflows
Retrieval-Augmented Generation system for querying documents.
A full-stack RAG application that processes uploaded PDFs, creates searchable representations, retrieves relevant context, and generates grounded answers from the source material.
Built with: React, FastAPI, PostgreSQL, PyMuPDF, embeddings, vector search
Agentic travel planning application.
An AI-powered travel planning system designed around agent workflows, conversational state, external information retrieval, and personalized itinerary generation.
Built with: LangGraph, LLMs, Tavily, MongoDB, Clerk