About
I build applied AI/ML systems. Started in 2014 on banking data engineering at TCS; the last 8+ years have been production ML — first at FedEx, then Starbucks, now co-founding VEZRAN.
What I focus on
Multi-agent systems and RAG pipelines that survive contact with production. Agent orchestration on frontier LLMs (Claude, GPT-4o, Gemini), eval design when ground truth is incomplete, and the unglamorous infrastructure — latency, cost, fallback paths — that decides whether agents actually ship.
What I’m building now
Co-founder and Head of AI/ML at VEZRAN. We’re shipping Zyberpol, an agentic SOC platform with four working agents — Correlation, Investigation, Triage, Remediation — sitting on top of CrowdStrike, Okta, Splunk, and AWS. The differentiator: signed, audit-ready evidence for every action. Pre-seed, hiring.
What I shipped before
- Starbucks — Senior Data Scientist, People Analytics (2022–2025). Turnover prediction and headcount forecasting models across the US retail footprint. BERT/T5 NLP on contact-center data: topic modeling, summarization, semantic search.
- FedEx — Senior Data Scientist (2018–2022). Package-at-risk and loss-intensity models in production on Azure across the logistics network. GDPR-compliant data matching at 95% retrieval. Fraud detection across shipping patterns.
- TCS — Systems / Data Engineer (2011–2014). Banking application engineering and automation work in Bengaluru.
Stack
Python, PyTorch, Transformers, LangChain / LangGraph, Azure ML, Databricks, MLflow, Ragas, Phoenix, Tableau.
Education
- MS, Business Analytics — The University of Texas at Dallas (2016–2017)
- BE, Information Science & Engineering — NMAM Institute of Technology, NITTE (2007–2011)