Case Studies & Architecture
A look under the hood at how I design production-grade Agentic systems.
Challenge: Automate the entire screening, matching, and shortlisting process without losing human nuance.
Solution: A LangGraph-orchestrated system with specialized agent nodes (Parser, Matcher, Ranker) backing onto a FastAPI + PostgreSQL/pgvector stack.
Challenge: Process high-velocity financial streams and extract actionable trading signals in under 5 seconds.
Solution: Dedicated WebSocket listeners feeding into an async processing queue, analyzed by parallel LLM workers with strict Pydantic output validation.
Projects
An AI-powered recruitment automation platform. Designed scalable AI-first system architectures and agent-based hiring workflows for rapid, high-quality talent acquisition.
An AI marketplace and app studio for discovering, creating, and monetizing AI assets. Engineered autonomous AI agents, multi-agent workflows, and prompt systems using LLMs and LangChain.
A financial intelligence engine processing real-time BSE/NSE data. Built a custom strategy builder and LLM-powered signal extraction system, reducing analysis time by 90%.
Real-time collaboration suite featuring AI-driven screen understanding. Engineered low-latency WebSocket infrastructure for seamless voice, video, and drawing synchronization.
An AI-powered analytics assistant transforming natural language into complex SQL queries. Delivers instant insights from a 300k+ row dataset of IPL match statistics.
GitHub Activity
My coding journey visualized — contributions and open-source work.