Building AI products · Open to part-time PM roles

Software Engineer → AI Product Manager

Nripesh
Kumar

UCLA Anderson MBA '28. Former Senior SWE at Visa. Building AI-native products at the intersection of enterprise fintech, LLMs, and human-centered design.

View Projects LinkedIn ↗

About

Engineer who thinks in products. Builder who ships.

4+
Years at Visa
2
Graduate Degrees
95
CareerLeader · Tech Interest

I spent four years as a Senior Software Engineer at Visa, helping launch V-ARM — Visa's first B2B accounts receivables product — and shipping an internal multi-agent AI tool that cut BRD review time from 2 hours to 2 minutes.

Before that, I earned an M.S. in Computer Science (ML specialization) from Columbia and a B.Tech. in CS from NIT Trichy. I'm now joining UCLA Anderson's MBA program to make the full transition into product management.

I'm currently building a portfolio of AI-native fintech tools — working in public, shipping fast, and documenting the tradeoffs. Available for part-time PM roles (20 hrs/week, remote) from July 2026.

Projects in progress

Building · Week 1–3

01

Payment Instruction Parser

LLM-powered extraction of structured payment data from unstructured emails — replacing brittle regex pipelines with a hallucination-aware AI layer.

Python FastAPI Claude API Streamlit PostgreSQL
PM signal: When does AI fail vs. regex? Documenting hallucination risk in financial contexts — a core AI PM interview answer.
Building · Week 1–3

02

PRD Quality Reviewer

Scores product requirements documents across clarity, completeness, edge-case coverage, and acceptance criteria quality — with section-level feedback.

Python Claude API Streamlit
PM signal: Shows I've read and contributed to real PRDs. I know what "good" looks like — rare for an engineering PM candidate.
Building · Week 4–7

03

B2B Payments Dispute Classifier

Classifies payment disputes by type, resolution path, and regulatory flag — with confidence thresholds that route low-confidence cases to human review.

Spring Boot Claude API PostgreSQL React
PM signal: AI confidence thresholds in compliance-adjacent contexts — how do you handle model error in a regulated industry?
Building · Week 4–7

04

SMB Financial Health Co-Pilot

Multi-doc RAG pipeline that ingests invoices and bank statements to surface cash flow risks, AR improvement suggestions, and cited insights for small business owners.

Python pgvector Claude API PostgreSQL
PM signal: User research + RAG tradeoffs + real SMB interviews. Bridges my consumer product experience gap.
Building · Week 4–7

05

API Reliability Monitor

Polls public APIs for latency spikes and error rate anomalies, then uses an LLM to generate plain-English incident reports and root cause hypotheses from raw logs.

Spring Boot Kafka Claude API PostgreSQL
PM signal: Platform/infra PM thinking — SLAs, developer experience, reliability. Strong fit for Stripe, Plaid, AWS PM roles.
Capstone · Month 2–3

06

AR Intelligence Platform

Full-stack AI product for mid-market finance teams. Multi-agent pipeline: overdue account tracker, payment likelihood predictor, and tone-aware follow-up email drafter. Subscription billing, auth, and real beta users.

Spring Boot pgvector Claude API React Clerk Stripe Vercel Railway
PM signal: The product I helped engineer at Visa — now owned end-to-end as a PM. Every AI tradeoff, user research session, and pricing decision documented.

Experience

2026 → 2028

UCLA Anderson

MBA, Full-Time

  • 705 GMAT Focus (750 GMAT). AnderTech, Management Consulting Association.
  • Concentrating on product management, AI strategy, and technology entrepreneurship.

Jul 2022 → Present

Visa

Senior Software Engineer

  • Co-defined MVP scope and GTM requirements for V-ARM — Visa's first B2B accounts receivables product.
  • Shipped a multi-agent BRD Scoring tool cutting manual review from ~2 hours to ~2 minutes.
  • Drove cross-continental delivery (Singapore/Austin), unblocking a 1-month delay and securing U.S. launch.
  • Championed generative AI for assistive coding, boosting deployment frequency by 25% and cutting time-to-quality by 30%.

Jun 2021 → May 2022

Columbia University

Graduate Research Assistant · E3B Dept.

  • Defined data acquisition requirements for a 2TB satellite imagery ETL pipeline, increasing throughput by 30%.
  • Translated researcher pain points into a self-serve imagery processing tool, accelerating rainforest mapping by 30%.

Oct 2020 → Dec 2020

Flatworld Solutions

Analyst Intern

  • Scoped and delivered an AI-powered document extraction solution for mortgage lenders, cutting processing time by 40%.

Technical fluency

LLM / GenAI Claude API Multi-agent Systems RAG Pipelines Python Java · Spring Boot Node.js React PostgreSQL pgvector Kafka Streamlit Vercel · Railway Agile / Scrum Product Roadmapping PRD Writing User Research A/B Testing

Get in touch

Let's build something that matters.

I'm open to part-time PM roles, fractional product work, and conversations with founders building at the intersection of AI and enterprise software. 20 hrs/week · Remote · Available from July 27, 2026.

nrikmr88@gmail.com LinkedIn ↗ GitHub ↗