Skip to content
Affan Shah

Affan Shah

AI engineer & full-stack developer

I build agentic AI systems — travel planners, market researchers, document assistants — and the full-stack products that make them usable.

Three projects, three kinds of system.

The archive holds 13 projects; 5 are live. These three show the range: retrieval, agents, and interface craft.

All 13 projects
001In build

TRIP PILOT AI

An AI-powered travel planning agent built with LangGraph, combining conversational interaction, tool-based research, and stateful workflow orchestration.

Python / LangGraph / LangChain / Ollama / Tavily

002In build

RAG PDF CHATBOT

An AI-powered document assistant that processes PDFs, retrieves relevant information using embeddings and vector search, and generates context-aware answers.

React / Python / FastAPI / PostgreSQL / PyMuPDF4LLM / Embeddings / Vector Search / JWT / Ollama / Groq

005Live

MOJITO

An immersive, scroll-driven 3D cocktail website featuring interactive flavors, realistic drink rendering, and cinematic animations.

React / Vite / Three.js / React Three Fiber / GSAP / Zustand / JavaScript / CSS

I build the retrieval layer, then the product around it.

Full-stack engineering with an applied AI focus. The interface is part of the system, not a wrapper added at the end.

I build full-stack products and the AI systems inside them. Most of my work sits in retrieval and agents: document pipelines, vector search, and LangGraph workflows that turn a model into something a user can actually operate.

On the product side I work in React, Next.js, TypeScript, and Python / FastAPI. I care about the unglamorous parts — state, errors, latency, and interfaces that stay legible when the model is wrong.

I am finishing a B.E. in Computer Engineering at St. Francis Institute of Technology in Mumbai, and going deeper into agentic architectures, retrieval quality, and production LLM tooling.

Current focus

  • Retrieval pipelines and evaluation for document-heavy products.
  • Agentic workflows with LangGraph, tool use, and state.
  • Full-stack product work in Next.js and FastAPI.

Available for work

Capabilities

Grouped by layer

Frontend
React / Next.js / TypeScript / JavaScript / Tailwind CSS / HTML / CSS
Backend & data
Python / FastAPI / Node.js / Flask / PostgreSQL / MongoDB
Machine learning
Scikit-learn / TensorFlow / Pandas / NumPy / Matplotlib
AI systems
LangChain / LangGraph / RAG pipelines / LLM integration / Vector search / Hugging Face

Education.

Coursework in computer engineering, applied to shipped projects.

2023 — 2027

B.E. Computer Engineering

St. Francis Institute of Technology, Borivali, Mumbai

In progress · Final year

Self-directed

Applied AI practice

Retrieval pipelines, LangGraph agents, and full-stack products built outside the syllabus.

Open the project archive