Final-year B.Tech CSE (IoT) student at VIT Vellore, focused on AI/ML, data engineering, and LLM-powered applications. I like building systems where the interesting part isn't the model, it's getting it to behave reliably inside a real pipeline, at scale, under real constraints.
Currently exploring: agentic workflows with LangGraph · retrieval-augmented systems · high-throughput data products
Full breakdown by category
|
A 10-node stateful agentic pipeline that automates enterprise complaint triage end-to-end - intent routing, entity extraction, and resolution workflows, replacing manual first-pass sorting.
|
Predicts re-order propensity across 1.3M+ transactions using XGBoost and CatBoost, cutting wasted outreach by ranking call-center targets by purchase likelihood.
|
|
High-throughput booking platform built for ticket-drop demand spikes - zero double-booking under concurrency, automated waitlist reallocation, live seat-map selection, and instant QR admission.
|
Streams real-time sensor telemetry over MQTT into a Random Forest model, delivering confidence-ranked crop recommendations from live field conditions.
|
Open to full-time roles, internships, and interesting builds - let's talk.