MotionArray
Product and project manager, leading the team through the company's acquisition by Artlist. I've sat on the buying side of a delivery relationship, which is why I estimate the way I do.
Someone or something else wrote it. Nobody left knows how it works. It still runs the business.
Right now
Founding engineer at BuildBetter.ai. An AI platform used by more than 20,000 teams that turns call recordings, tickets and support conversations into structured, cited documents. My work there is the unglamorous half of shipping AI: retrieval over a company's own material so answers are grounded instead of plausible, structured extraction from input that arrives messy, and the evaluation and cost work that decides whether a feature survives real usage.
I own billing, including the top-up model, along with the Slack, Calendar and Intercom integrations and the call ingestion pipeline. That's the perimeter of the product: every place money or data crosses in or out. It also means shipping a new surface one week and optimizing the database behind a feature I've never touched the next, which is what founding engineer means at a company still small enough for the title to be accurate.
Built Ledgely from zero. A product for agencies that pulls timelogs out of Toggl, Harvest, Hubstaff, Upwork and Jira into one view, watches fixed price budgets against pace, and raises the alarm before the margin is gone. Schema, integrations, billing and interface are mine. I built it because every project I ever lost money on was fixed price, and I never saw it coming until the invoice didn't add up.
Founder of Roraima Devs. I'm Venezuelan, I run the company from Florida, and it's a U.S. entity. The team of 30+ engineers across Venezuela, Colombia, Argentina, the Dominican Republic and the U.S. isn't an outsourcing arrangement I assembled from a directory. It's a network I'm actually from, which is why the engineers stay and why clients pass the three year mark. I still write code every week, and when a project needs frontend, backend, mobile and infrastructure moving at once, I can staff all four in parallel instead of stretching one timeline into six months.
Strata
Every product I inherit has layers: the framework it was born in, the rewrite that stopped halfway, the integration bolted on under deadline. You can date each one. Reading those layers correctly is the difference between a migration that ships and a rebuild that stalls.
Dated by when the code was written, not when it reached me
The MVP worked, the demo closed the round, and now nobody on the team can say why it breaks at load. A codebase can be legacy at six months old if no one left understands it. Same job as a twenty year old system: read what's there, find what's load bearing, change it without stopping the business.
Retrieval grounded in a company's own documents, structured extraction from unstructured input, and the evaluation work that tells you whether the output is good enough to charge for. Most products don't need a model. They need the right context reaching one.
Fitness and marketplace products built as the frontend lead inside another agency's delivery team, taken from design files to app stores on someone else's brand.
A .NET codebase older than most of the frameworks brought in to replace it, migrated in stages while it kept processing transactions every day. No freeze, no big-bang cutover, no month where the business ran degraded so the rewrite could land.
Frameworks past end of life, no documentation, no original team, and a business that depends on them running Monday morning. One of these had no version control at all when it reached me. The first commit in its history is mine, written years after the system went into production. This is the work I get asked for most, and the work most developers quietly decline.
SurfaceBedrock
Tools
I pick by what your system already runs on, not by what I'd enjoy writing.
Before
Product and project manager, leading the team through the company's acquisition by Artlist. I've sat on the buying side of a delivery relationship, which is why I estimate the way I do.
Founding team and tech lead at a startup building sustainability reporting. Early architecture, early hires, and the specific discipline of choosing what not to build yet.
Start with a 30 minute call. I'll ask questions before I estimate, and if your idea has a cheaper path than the one you arrived with, I'll say so on that call. If I'm not the right fit, I'll say that too.