I build the core plumbing for high-scale, AI-powered data platforms

Product Manager @Visa · Ex- NetApp, Walmart, CommScope, SIXT

Visa Flipkart Walmart NetApp CommScope SIXT
4.2B+
Data Points (network scale)
50+
Global Clients (implementation success)
92%
Adoption (within 6 mo)

My Core Competencies

  • Data Infrastructure
  • Product Strategy
  • Cross-functional Leadership
  • AI/ML Products
  • Platform Engineering
  • Fraud & Security

Why I Think Differently (Engineer-to-PM DNA)

I meet at the intersection where engineer-to-PM DNA converges — I can read a data schema, challenge an architecture tradeoff, and build to production-grade standards.

Verify on GitHub
Vijeta Bhatia
"Production AI is not a demo problem, it is a data infrastructure problem."
Vijeta Bhatia

Vijeta Bhatia is a product leader building technical products for data infrastructure, fraud systems, and AI platforms in the Bay Area.

I started on the engineering side — building Flipkart's first payment gateway for 500K+ daily transactions. Five years of writing code, debugging fraud rules, and watching what broke taught me one thing: the most valuable product insight is always hiding in data no one thought to collect.

At Visa, that instinct converged with scale. I owned the fraud data platform processing 4.2 billion annual transactions, found a blind spot covering 35% of all transaction volume, turned it into a global signal — then shipped agentic AI on top of it.

Read more...
Built on Visa Infrastructure

AI Lab: Pioneering Production AI Readiness

I experiment with frontier AI techniques and build production-grade systems that solve real business problems.

AI Agent for Fraud Investigation

Multi-turn conversational agent for fraud analysts. Combines RAG over transaction graphs, real-time data access, and agentic reasoning to accelerate investigation velocity.

View showcase →

Autonomous Anomaly Detection

Unsupervised learning system for discovering novel fraud patterns. Designed for low false-positive rates and cross-merchant generalization in payment fraud.

View showcase →

0→1 Declined Auth Data Signal

The foundational platform that powers all downstream AI. Billions of declined messages, engineered into a first-class global payment signal.

View showcase →

Production AI Readiness Evaluator

Self-assessment framework for evaluating whether an AI system is production-ready. Covers data quality, inference latency, explainability, and governance checks.

Launch Evaluator →

Selected Projects

Declined Auth Data Platform

Discovered that 35% of all transaction volume (declined authorizations) was missing from ML pipelines. Designed the data model, built the ingestion pipeline, and launched it as a first-class global payment signal — expanding ML coverage from 65% → 94%.

View deep dive →

Agentic AI on Fraud Data Platform

Conceived, scoped, and shipped an agentic AI layer (RAG + LLM orchestration) on top of Visa's fraud data platform — from whiteboard to production. Reducing investigation time by ~30%.

View deep dive →

Autonomous Anomaly Detection

Built an autonomous anomaly detection system on the fraud data platform — from 68% false positive rate to ~35%, with confidence-gated automation that auto-resolves clear patterns.

View deep dive →

FRECOP — Global Fraud Data Standard

Scaled Visa's Fraud Reporting and Control Program from a pilot to a global platform standard across 50+ issuers and acquirers. Drove 92% adoption in 6 months.

View deep dive →

Experience

Visa
July 2023 – Present
Product Manager – Global Data Team (Global Risk & Fraud)
Visa's fraud data platform is the connective tissue of global payments. My job was to own it — every data flow, every transformation rule, every ML signal.
  • Owned Visa's core fraud and transaction data platform end-to-end, supporting 4.2B+ annual transactions across 50+ global issuers and acquirers.
  • Led 0→1 launch of declined authorization data as a first-class payment signal — expanding signal coverage from 65% → 94%.
  • Conceived, scoped, and shipped agent-assisted AI (RAG + LLM orchestration) — reducing investigation time by ~30%.
  • Scaled Visa's FRECOP into a global platform standard — 92% adoption in 6 months, 75% fewer P1 data incidents.
  • Led a 30+ person cross-functional team; mentored 3 junior PMs.
NetApp
May 2022 – April 2023
Product Manager – Product Security / StorageGRID
First PM role in the US after MBA — defining product strategy for a security analytics platform.
  • Led product discovery across 10+ internal teams to define requirements for secure, compliance-aware analytics platforms.
  • Designed executive-facing security dashboards that shaped roadmap priorities.
  • Increased internal adoption of security tooling by 60% through A/B-tested UX improvements.
NC State — Full-Time MBA
Aug 2021 – May 2023
STEM MBA — Innovation, Marketing, Management & Business Analytics
CommScope / Ruckus
Sept 2020 – July 2021
Principal Software Engineer / Technical Lead
  • Drove requirements, architecture tradeoffs, and release management for enterprise Wi-Fi and cloud network management.
  • Led cross-geo platform ownership transfer — built and led a 10+ person engineering team.
SIXT Research and Development
April 2018 – May 2020
Product Owner / Lead Engineer
Directi
March 2016 – March 2018
Product Quality Strategist
Flipkart / Walmart
Jan 2013 – Feb 2016
Software Development Engineer
Core contributor to Flipkart's first payment gateway (Pay Zippy), processing 500K+ daily transactions.

My Playbook

Let's build something together.

Senior PM / Staff PM / Group PM · Fintech, payments, AI platforms · Growth-stage or big tech

I've spent 13 years building platforms at scale. I ship AI systems, lead cross-functional teams, and turn data into decisions. If you're building fintech, payments, or AI infrastructure — let's talk.