I'm a final-year B.Tech student in AI & Machine Learning at BIT Mesra (CGPA: 8.7), building production-style systems that combine backend engineering, applied ML, and data storytelling — not just notebooks.

I've interned at two companies building real systems — an ERP procurement automation platform at AxisPoints, and a RAG-based document QA pipeline at Mecon Ltd. Since then I've been building larger, end-to-end projects on my own: multi-agent AI orchestration, a fairness-aware credit underwriting engine, and NL-to-SQL systems grounded in RAG.

Currently on the lookout for SDE and Data roles.

Languages & Tools

Backend & Systems

AI / ML / Agentic Systems

Data & Analytics

Production-style agentic system where a Planner, Research, Database, Code, and Synthesis agent collaborate to complete complex requests. Built on LangGraph for orchestration with human-in-the-loop approval checkpoints, step/budget limits, and loop detection. LangGraph FastAPI PostgreSQL Redis OpenRouter

End-to-end loan underwriting pipeline replicating a real bank's decisioning flow: a deterministic rule engine, a fairness-constrained LightGBM model (via fairlearn), per-applicant SHAP explanations, a manual review queue with RBAC, immutable audit logging, and live drift monitoring with Evidently AI. FastAPI LightGBM SHAP Fairlearn MLflow Evidently AI Docker

Procurement analytics platform combining a RAG-powered natural language query interface over spend data with statistical anomaly detection (duplicate invoices, vendor spend concentration, contract expiry alerts) and an optional Odoo ERP connector. FastAPI LangChain FAISS React PostgreSQL Odoo XML-RPC

Natural language–to–SQL engine grounded in a dual RAG pipeline (schema + business glossary), with a SQL validation layer, confidence scoring, and auto-generated visualizations for query results. FastAPI Groq (LLaMA 3) FAISS PostgreSQL Streamlit

Open to SDE, Data Analyst, and Data Scientist opportunities 👀