I design autonomous, agentic workflows that turn LLMs and APIs into software that solves real business problems — built with Python, reasoning models, and a product engineer's mindset.
I'm an early-career engineer, but I don't build like one. I'm less interested in collecting frameworks and more interested in understanding how autonomous systems actually make decisions — how an agent reasons, when it should call a tool, and how to keep it reliable in the real world.
Most of my time goes into a simple question: what work here shouldn't be done by a human? Then I connect the right models and APIs into a workflow that handles it end to end — triaging inbox chaos, answering with voice, routing decisions to the people who need them.
I care about outcomes. A project only matters if it saves someone time, reduces a delay, or removes a decision that used to require effort. That's the mindset I want to bring to a team building production AI.
If a task is manual, predictable, and draining, I see a system waiting to be built.
Real value comes from connecting APIs and models into flows that run on their own.
I care about reasoning, tool use, and context — not just prompts that happen to work.
I chase working products that solve problems, not tutorials I never finish.
Each project started with a real problem and ended with software that runs on its own. Expand any card to see the reasoning behind it.
An AI system that prioritizes leads, filters noise, and helps sales teams focus on conversations that actually matter.
The workflow
A voice-first assistant that holds natural conversation and takes action on your system — hands free.
The workflow
I don't collect tools — I connect them. Hover or tap each layer to see how raw code becomes real business value.
The foundation
Where every system starts — the language I use to wire logic, models, and services together.
Organized the way I actually use them — from the language and the models to the APIs and libraries that turn ideas into running automation.
Good AI products aren't magic — they're the result of understanding a problem deeply and refusing to ship until it genuinely works.
Before any code, I find the real friction — the task that drains time or delays a decision.
I map the flow first: what triggers it, what decisions happen, and where the value lands.
I add LLM reasoning exactly where judgement is needed — never more, never for show.
The system handles the loop end to end so humans only touch what genuinely needs them.
It only counts when it runs, saves time, and someone actually relies on it.
Learned to think in logic and structure — the foundation everything else builds on.
Discovered that the real power is connecting services, not building everything alone.
Started designing systems that reason and act, not scripts that only follow rules.
Turned ideas into working automation — Inbox Intelligence and Jarvis among them.
Ready to bring this mindset to a team building ambitious AI products at scale.
I'm looking to contribute to ambitious AI teams building products that create real-world value. If that's what you're building, I'd love to talk.