I build enterprise AI products. I prototype them, define how we measure success, and ship. Here's proof.
I take ambiguous enterprise problems, prototype the first working version, define the evaluation framework, and ship.
Talent intelligence, job market analytics, dynamic pricing, document intelligence. Products that work at scale in regulated environments.
Evaluation frameworks, KPI trees, human-review workflows, false-positive tracking. I define how we know it works before it ships.
Discovery, PRDs, user stories, acceptance criteria, UAT, release planning. Ambiguous requirements to shipped product.
I prototype with Claude Code, build RAG applications, design evaluation harnesses, and ship working AI tools.
Confirmed outcomes labeled as such. Designed targets labeled as such.
AI-driven employment profiling, skills taxonomy, and job market analytics serving a national government workforce agency.
AI-powered internal mobility, skill-based matching, and career pathing for a top-10 global HCM SaaS platform.
Reinforcement learning-driven pricing optimization from PRD through measured pilot outcomes.
0-to-1 B2B SaaS for managed service providers: ingestion, normalization, audit reporting, telemetry.
Each build solves a real problem and documents the product decisions behind it.
Production-style RAG application. Upload any PDF, ask questions, get cited answers. Every architecture decision documented as a product tradeoff.
Why: A PM who says "we chose MMR over cosine similarity because our docs repeat" beats one who says "the AI retrieves content."
AI-powered home construction companion. Track progress, manage contractor milestones, estimate costs, make informed decisions through the build journey.
Why: Construction is India's most broken consumer experience. After my PropTech PM work, I knew exactly where the pain was.
Try itInteractive framework generator. Pick an AI product type, get tailored metrics, failure modes, human-review triggers, and go/no-go criteria.
Why: Most AI products fail because nobody defined "good" before shipping. This tool forces that conversation.
Try it15-question assessment across 5 dimensions. Get a radar chart score and actionable recommendations for your AI readiness.
Why: Every enterprise wants AI. Most aren't ready. This replaces the vague consulting assessment with an honest diagnostic.
Try itEverything listed maps to shipped work.
Frameworks and evaluation patterns for AI product decisions.
Building something complex? I'd like to hear about it.
India · Middle East · US