Case Studies · 4
Selected work in AI product.
Real problems, the trade-offs behind each decision, and the outcomes that followed — across GenAI, ML, and shipped product.
- 01 · Workhuman · 2025
Workhuman iQ — Turning peer recognition into workforce intelligence
An enterprise AI platform that reads millions of recognition signals and surfaces who to grow, who to retain, and where culture is moving.
5AI surfaces shippedEnd-to-endResponsible-AIProduct leadRole - 02 · Laya Healthcare · 2024
Laya Claims AI — From 14-day claims to same-day reimbursement
Computer vision + OCR pipeline embedded in the member app that reads a claim, scores it, and routes it in under 60 seconds.
In-appChannelOCR + CVPipelineContract leadEngagement - 03 · AMCS Group · 2023–2024
AMCS GenAI — Conversational ERP that real operators use
A GenAI interface on top of an enterprise ERP that lets dispatchers and operators ask in plain English what used to take three reports and a phone call.
30d → 3sProcess re-engineeredWinner '22 & '23Innovation AwardGenAI Product OwnerRole - 04 · EGlobetech India · 2017
EGlobetech EMG — Teaching a bionic limb to read intention
An early R&D project: training ML models on EMG signals so a bionic limb could distinguish one limb movement from another — the first model I shipped, and the start of the constraint-thinking that runs through everything after.
Pure MLDisciplineR + PythonStackR&D InternRole