🌐 Live portfolio → himanisharrma.vercel.app
Product manager for AI-native products in fintech — I own the arc from discovery and strategy through roadmap, architecture, and launch. Rare for a PM, I also build hands-on. Most recently I took KYC, Re-KYC & compliance platforms from zero to shipped with agentic AI tooling.
- 🎯 I own AI-native products end-to-end in fintech — discovery, strategy, roadmap, and delivery — turning ambiguous problems into shipped software
- 🤖 A PM who can build: I prototype and ship hands-on across a broad, multi-vendor AI stack — coding agents, Cowork, MCP, RAG/evals, and more (full toolkit below) — so ideas become working products, not just decks
- 🏦 Fintech across KYC/RegTech, payments ops & GTM — most recent depth: RBI KYC · V-CIP · Re-KYC · maker-checker · audit & PII
- 📐 How I think about AI in products: AI assists · rules decide · humans approve · logs prove
- 🌱 Currently exploring: multi-agent orchestration — role-specialized agent pods, clean hand-offs, and reliability beyond a single agent
- 📫 Reach me at himani1666@gmail.com — full case studies & diagrams live in my portfolio hub
Engineering stack
How I build with AI
My AI-native toolkit
| Layer | Tools |
|---|---|
| 🧠 LLMs & research | Claude · ChatGPT · Gemini · Perplexity · Grok · NotebookLM · M365 Copilot · open-weights (Llama / Mistral / DeepSeek) |
| 💻 Coding & agents | Claude Code · Cowork · Cursor · Windsurf · GitHub Copilot · OpenAI Codex · Replit · Devin · Claude computer-use · OpenAI Operator · Browser Use |
| ⚡ Build & prototype | Bolt · Lovable · v0 · Google AI Studio · Figma Make · Galileo AI · Uizard · Framer AI · Remotion |
| 🔗 Agent engineering | MCP (I build servers) · RAG + evals · structured outputs · prompt caching · LangChain · CrewAI · Flowise · n8n · Zapier · Make · Supabase |
| 📋 PM · docs · design | Notion AI · Jira · Linear · Confluence (Rovo) · Productboard · ChatPRD · Figma · FigJam / Miro · Whimsical · Napkin AI · Mermaid |
| 📊 Data · notes · research | ChatGPT Advanced Data Analysis · Claude analysis tool · vision-OCR · Otter · Fireflies |
| 🎨 Multimodal & cloud | AWS Rekognition / Textract / Comprehend · Midjourney / DALL·E / Ideogram · Runway / Pika / Kling / Veo · HeyGen / Synthesia · ElevenLabs |
Fintech, built end-to-end with AI — spanning KYC/RegTech, payments ops, and GTM automation.
| Project | What it is | Highlight |
|---|---|---|
| Identity-Verification-Amazon-Rekognition | Extended AWS's Rekognition identity-verification sample into a full RBI Video-KYC platform — WebRTC liveness + Maker-Checker on a 32-migration schema | 141K LOC · ~3,000 tests |
| rekyc-compliance-module | Periodic Re-KYC engine + eval-measured RAG over the RBI KYC Master Direction (BM25, prompt caching) | 34K LOC · 12/12 evals |
| video-kyc-hackathon | Video KYC app + a 6-tool MCP server with PII masking & human-gated verdicts | 9K LOC · 6-tool MCP |
| re-kyc-with-landing-page | Full-stack multi-role KYC platform with bulk Excel-driven Re-KYC triggers | 59K LOC · 5-role |
| finverge-sales-skills | Claude Code plugin automating sales collateral via a permission-aware dual knowledge base | 17K LOC · 8 skills |
| payops-copilot | Evidence-first payment-reconciliation workspace | 11K LOC · recon |
🗂️ All case studies, architecture diagrams & my AI toolkit live in the portfolio hub →
| I design | I enforce | I measure |
|---|---|---|
| Agent workflows, context engineering, tool surfaces (MCP), FSM-driven lifecycles | Hooks & guardrails, PII masking at boundaries, human-in-the-loop gates, audit trails | Eval suites before shipping prompts, verification via browser agents, coverage targets |
300+ commits · ~300K LOC across public repos · a 6-tool MCP server · a 12-case eval suite over a grounded RAG pipeline · an 8-skill Claude Code plugin · a 32-migration production schema · ~3,000 automated tests
💬 Ask me about: AI-native product management · agentic product development · fintech (payments, KYC, RegTech)