I'm a Computer Science student who learns by shipping. Most of what I build sits at the intersection of LLM agents and real systems — tools that plan, retrieve, and act rather than just autocomplete.
- 🔬 Deep in agentic workflows, RAG pipelines, and multi-agent orchestration — Corrective RAG and Self-RAG implementations, LangChain/LangGraph, plus an agentic platform built for a hackathon.
- 🦀 Also drawn to the layer underneath: a packet-filtering firewall in Rust and a network intrusion detection project keep me honest about what's actually happening on the wire.
- 🚀 I ship end-to-end — TypeScript/Next.js front to back, Python for the ML side, deployed and publicly reachable.
- 🤝 Open to collaborating on AI tooling, developer experience, and anything agentic.
Languages
Frameworks & Runtime
AI / Data
Tooling & Infra
LangChain · LangGraph · Clerk · Liveblocks · Trigger.dev · React Native · Jupyter
|
Intelligence-led police investigation platform. Agentic AI covering the full case lifecycle: multilingual complaint ingestion (Gujarati PDFs, Hindi audio, handwritten FIR scans), SOP-grounded investigation paths, automated legal drafting, and audit-transparent case summaries. Built for the ERH26 Hackathon.
|
Interactive systems-architecture builder. Design and reason about system diagrams collaboratively, with real-time presence and background job orchestration.
|
|
Financial analysis and insight engine. Python pipeline for ingesting market data and surfacing signal — the analytical sibling to my algorithmic-trading experiments.
|
AI assistant for Git workflows. Reduces the friction between intent and the right command — built because I kept reaching for the same handful of incantations.
|
|
Packet-filtering firewall written in Rust. A deliberate step down the stack: rule evaluation and packet inspection without a garbage collector in the way.
|
Homework automation agent. Point it at an assignment and it works the problem end-to-end. Equal parts useful tool and study of agent reliability.
|
📂 More projects — AI research, trading, and tooling (click to expand)
Agentic AI & RAG
| Project | What it is | Stack |
|---|---|---|
jobpilot |
AI job-search agent — role matching, tailored resumes, faster applications | TypeScript |
agent |
Agent scaffolding and experiments | TypeScript |
crag |
Corrective RAG implementation | Python |
selfrag |
Self-RAG — retrieval with self-critique | Python |
chatbot |
Conversational agent groundwork | Python |
lagnchain-langgraph-tutorials |
LangChain / LangGraph working notes | Jupyter |
Products & Platforms
| Project | What it is | Stack |
|---|---|---|
Command-Bot |
Command-driven bot — live ▸ | TypeScript |
odoo-hackathon |
TransitOps — smart transport operations platform | TypeScript |
vocaba |
Vocaba/Lingua — language learning | React Native Clerk |
nodebase |
Node-based workspace experiment | TypeScript |
Quant & ML
| Project | What it is | Stack |
|---|---|---|
algotrading |
Algorithmic trading strategies | Python |
algoX |
Algorithm visualization and practice | TypeScript |
DSlab_Project_Network_Intrusion_detection... |
Network intrusion detection — Data Science lab | Jupyter |
🏙️ See my contributions as a 3D city (click to expand)
