Headline

Data Scientist | LangGraph · Agentic AI · CV | RAG Systems · Object Detection · MCP Protocol

About

I build AI systems that actually ship — not Jupyter notebooks. At Incedo, I engineered a production-grade Computer Vision system using YOLOv11 that detects squash ball and racket positions across 20,000+ frames with deployed to real clients, with monitoring, logging, and containerised delivery on Azure. On the GenAI side, I've architected modular Agentic RAG pipelines using LangGraph, with reusable nodes (Retriever, Rewriter, Grader, Generator) and JSON config-driven orchestration. I also built a Jira automation agent using the Model Context Protocol (MCP) — one of the newer integration patterns very few engineers at my level have worked with in production. What I bring that's less common at ~2 years of experience: 1. End-to-end ownership: dataset creation → training 2. deployment → monitoring 3. Agentic AI architecture, not just prompt engineering 4. Infra thinking: Docker, FastAPI, Redis job queues, Azure ML I'm currently exploring Senior Data Scientist and AI/ML Engineer roles where I can work on hard problems at the intersection of LLMs and real-world systems. Open to conversations — feel free to reach out or connect.

Location

Delhi

Work experience

Data Scientist at Incedo Inc., Aug 2024 - Present · 1 yr 9 mos, • Engineered a production-grade Squash Referee AI using YOLOv11 — detecting ball and racket positions across 20,000+ annotated frames ; deployed to client with full monitoring, logging, and containerised delivery on Azure ML. • Architected a modular Agentic RAG system using LangGraph with pluggable nodes (Retriever, Rewriter, Grader, Generator), JSON config adapters, and mock node support — enabling rapid offline testing and configuration-driven LLM workflows at scale. • Built a Jira automation agent using Model Context Protocol (MCP) with LLM-based intent parsing, field mapping, and priority classification — reducing ticket creation time from minutes to seconds via a FastAPI chat interface. • Delivered containerised, production-ready AI services with integrated validation, observability, and client deployment pipelines across multiple engagements.

Application Analyst Intern at Accenture India, Jun 2023 - Jul 2023 · 2 mos, Streamlined incident workflows using Trend Micro Deep Security Manager. Implemented UI Path automation to eliminate repetitive manual tasks. Built a React.js + Tailwind dashboard for real-time incident tracking.

Education

Bachelor of Technology - BTech, Computer Software Engineering, Delhi Technological University (Formerly DCE)

Senior Secondary, Rishabh Public School - India

Public links

github: https://github.com/vishalpathakcodes