Product manager · Fairfax, Virginia
Venkatesh Gardas
I turn messy operational work into software people open on a Monday morning.
I turn messy operational work into software people open on a Monday morning.
An AI risk platform for audit teams. Decision cycles down 16 percent, adoption at 90 percent.
Job cards, customer approvals and GST invoicing, working at a real counter every day.
Voice-first AI for elderly parents, set up by family from another continent.
Building PitStop with a live garage, running the Saathi beta with real families, and looking for a product role.
Garage management for Indian workshops. Job cards, customer approvals, GST invoicing and UPI, built for a service advisor with oily hands and a queue at the counter.
See inside →A voice companion for elderly Indian users. It speaks Malayalam and Telugu, the type is enormous, and the family sets it up from another continent.
See inside →Home security hardware for the Indian market. An alarm panel, sensors and an app, priced for a family that has never bought a security system before.
See inside →An AI app that tears a resume apart and hands back honest, useful feedback at three roast levels. Built and shipped in a weekend, which is its own kind of scope lesson.
Try it live →Six years across data, fintech and product, three products of my own, and one of them earning its keep at a real counter.
Five jobs, one direction. I started fixing other people’s code and ended up deciding what gets written.
I own the full lifecycle of an AI risk analytics platform: 40 plus automated risk decision flows, decision cycles down 16 percent, adoption at 90 percent.
What I do there → ProcurementB2B procurement for clients like Bajaj, HDB Hiranandani and Akasa Air. Rebuilt the budget module and built a loan origination system that shipped as a paid add-on.
What I did there →SQL pipelines, cohort models, price determination and catalog modeling for purplle.com. That work drove a 12 percent drop in churn and 15 percent growth in repeat orders.
Large ERP and logistics modules: pricing for manufactured products, requirement analysis, the website, and every piece of analysis in between. Delivery risk down 25 percent.
JavaScript and Python, other people’s bugs, users on the phone while you fix them. The best possible school for a future product manager.
Each step was a deliberate trade: support to get closer to users, analysis to get closer to decisions, data to prove them, product to own them.
The master’s that brought me to the US. Finished May 2026, while shipping three products of my own on the side.
Earned while working full time as an analyst and then a data modeler. The theory and the practice arrived in the same week, repeatedly.
Where the code started.
Everything useful I have shipped came from standing where the work happens and watching what people do instead of what they say they do.
Data model, permissions, the ugly edge cases. Get those right and the rest of the roadmap stops fighting you six months later.
A feature nobody opens is a feature that does not exist. I care about the second week of usage far more than the launch.
Streets, machines and light, on a Sony α7 V. Thirty one frames and counting.
See the gallery →Weekends smell like engine oil. Every screen in PitStop exists because I have stood at that counter and watched the diary lose a job card.
A machine in the corner of the house runs Jellyfin behind a reverse proxy, with certificates that renew themselves. Mostly. The 2am debugging is how I learned what my engineers deal with.
I am looking for a product management role in vertical SaaS, fintech or AI tooling, in the US or remote. Authorized to work in the US on OPT. Happy to talk shop about garages, approval matrices or alarm panels too.
venkatesh@venkateshgardas.com