I’m Robert — a full-stack engineer in Chicago building backends in Python/FastAPI, frontends in React and Vue, and AI-agentic systems that hold up outside a demo. I find the problem worth solving, then build the thing that solves it.
Finding the real gap before building
Live PoC
Shipped to both founders
Problem
The team had a working app but no signal on where analysis effort was actually paying off for users — feature calls were being made on instinct.
Approach
Audited the app and ran UX interviews with the founding team before writing code. Built a full-stack PoC targeting the real gap, then presented it for a go/no-go.
Automating the demo bottleneck
-20%
Manual production time
Problem
A white-label email-marketing pipeline needed PoC demos built fast enough to keep sales moving — manual production was the bottleneck.
Approach
Built an agentic system on Hermes OS, combining ElevenLabs voice AI and GPT-based video generation to automate PoC creation, prototyping, and demo recording end to end.
Research before the rewrite
-10%
Page load time
Problem
Two platforms serving 5,500 combined users were carrying slow page loads — nobody had validated what was actually driving the friction.
Approach
Interviewed and hired a UX researcher to audit the platform first. Her findings shaped the case for and scope of a JS-to-HTMX migration across both platforms.
The Quorum + Loop
This is the real method behind the case studies above: run the question past three differently-profiled advisors — an engineer, a business owner, and the stakeholder living with the result — then have an agentic critique pass (“Loop”) stress-test the synthesis before the advisors get a last honest word. Costs are capped per submission, so this may occasionally skip a low-priority step to stay within budget.
Or see a saved run instantly — no cost, no captcha