
FitWardrobe
An AI wardrobe stylist that turns a closet full of clothes into daily outfits people actually wear.
A Product Evidence Book | AI Product Manager
I'm Aryan Panwar. I'm an AI Product Manager with a builder background. This isn't a portfolio: it's an evidence book. Case studies from agentic systems I've shipped, the tradeoffs I made, and the writing that shows how I think about product.
PM toolkit
Technical fluency

02 · Flagship Case Studies
Problem, users, tradeoffs, metrics, and what I'd do next - for the three projects I've thought about hardest.

An AI wardrobe stylist that turns a closet full of clothes into daily outfits people actually wear.

An open-source voice AI toolkit that lets small teams ship real-time voice agents without a research team.

Stop AI IDEs giving stale SEO advice - a 14-phase agentic skill for Cursor, Windsurf, and Copilot.
Supporting case studies and the full archive live on the Case Studies page.
03 · How I Think
Step
Sit in the problem. Interview users, read support tickets, watch session replays. Assume nothing.
04 · My Product Thinking
Every entry starts with a product I shipped - not with theory. JTBD, RICE scoring, opportunity trees, and metrics - applied to real products.
05 · Writing
Each post answers one PM hiring question - grounded in a real product I built, not a framework I memorised.

FitWardrobe started as a computer-vision demo. It only became a product when I stopped defending my prototype and started listening to the seven friends who told me it solved the wrong problem.
Answers: How do I discover the right problem?

The moment my technical instinct stopped being an asset and started being a tax - and the three questions I now ask before I let myself open a code editor.
Answers: How did my mindset evolve from builder to PM?

The one-page pre-PRD I run before any spec - and why the "must-have" column is the one that decides whether the product ships on time.
Answers: How do I decide what gets built?

The Credex feature I almost shipped without a definition of "worked" - and the two-column doc I now use so I never do that again.
Answers: How do I know a feature is successful?
06 · Journey
The milestones - hardware, research, shipped AI products, and where I'm heading next.
2022
Enrolled in Electronics & Communication Engineering. Spent the first year building hardware side-projects - a robot arm, an EV-charger prototype, staircase lighting - for the reps of finishing things.
2023
Discovered that the interesting problems live at the intersection of user, business, and technology. Devoured Marty Cagan, Teresa Torres, and Julie Zhuo. Started writing PRDs for products I was only imagining.
2024
Stacked reps on hardware side-projects - a 4-DOF robotic arm, an IPT wireless-charging prototype, motion-triggered staircase lighting. Every build had to actually work, measure something, and be handed to a user. That constraint discipline is the seed of how I now scope products.
Jun–Aug 2025
Spent 10 weeks inside IPL's molasses distillery in Rohana Kalan, Muzaffarnagar - learned how industrial automation systems (PLC/DCS, boilers, sensor arrays) operate across full production. Developed first-principles respect for safety compliance and enterprise operational constraints.
Dec 2025
Went from side-projects to a real shipping cadence: FitWardrobe (on-device AI, live), AI-First HCP CRM on LangGraph, Credex AI Spend Audit (live), Mithivoices open-source voice AI, and Digital Heroes Golf (live with payments). Learned that scope discipline is the whole game - and that AI wedges only work when the ritual around them is designed.
Feb–Mar 2026
Earned NVIDIA DLI's Transformer-based NLP certificate (Feb), Google AI Essentials (Feb) and Prompting Essentials (Mar), and Microsoft's Product Management Fundamentals (Mar). The PM track stopped being a side interest and became the direction.
April 2026
‘Automated Dual-Spot Wireless Charging and Parking Management System for EVs’ published in JETIR (IF 7.95, Vol 13 · Issue 4, ID JETIR2604872, Apr 27 2026) - 85% wireless charging efficiency, 1st place at the MIET branch expo. The paper's structure - assumption, method, result, discussion - quietly became how I write PRDs.
Read Published Paper (JETIR ↗)2026
Published an npm package that makes sites legible to AI answer engines (Google AI Overviews, ChatGPT, Perplexity). Shipped 11+ essays on positioning AI products, orchestrating LLMs in production, and designing for agentic users.
View npm Package ↗Next
Looking for a product team where the bar is high, the users are real, and the outcome matters. I want to spend my early PM years in the room where the tradeoffs are made.
07 · Product Philosophy
“The PM job isn't to have the best ideas. It's to make sure the team ships the right ones, for the right users, with the right restraint.”
01
02
03
04
08 · Let's Talk
I'm looking for a product team where the bar is high and the users are real. If that sounds like you, get in touch.