Shipped, open source
Perfect Placement
Your job search, run by your agents, on your machine. It finds roles, scores fit, tailors materials and drafts outreach. It never sends or submits without you.
Try it on GitHubHumans decide. Agents carry.
I work inside products that already exist: your code, your customers, your team. Anyone can spin up a new app in an afternoon. I make the one you have better, fast, and show your team how as we go.
Try it below. The agents carry, you make the calls.
Agent carriesYou decideWorking in: your existing product
What did your customers just tell you?
Your call: which job matters most?
Your call: send the proof to the customers who raised it?
The proof, inside your app
Coming soon to this screenYour next improvement
Pick what your customers said and watch the change land here.
An illustration of how I work. The scenarios are made up.
Where I've built
G-P, Atomic, Portico (CourseKey), Rosalyn.ai and The VL Studios.
Off the anvil
Specdr
Ongoing, with the Specdr team
What comes after agile, for products that already exist. Proof-driven development: prove the problem, prove the solution, prove the impact. Customer signal becomes a working proof inside the real product, goes back to the customers who asked for it, and reaches engineering as evidence instead of a ticket.
Read more at specdr.aiRIP'D
Sunset
A trading card platform for collectors, with computer vision and AI card recognition. 40 organic daily users within three months of launch.
Read the case studyPandora
In the forge
A self-hosted second brain: a personal context layer with curated agents on top. Open source is on the roadmap.
Have a product? Let's make it better.
I start from what you have. Early-stage, growth or enterprise, the work happens inside your product and alongside your team, so they can keep running it after I'm gone.
Proof Sprint
One week inside your product: prove the problem, then put a working solution in front of the customers who raised it.
AI Product Foundations
The evals, retrieval and data layer your existing product needs to ship AI features that hold up.
Fractional AI PM
A player-coach on your team, running the loop on your product and teaching your team to run it without me.
Founding PM at Atomic, where we raised an $800K pre-seed. Principal PM for the data and AI platform at G-P. Advised three early-stage AI startups; two raised over $3M.
Book an AI Readiness ChatMoving a whole engineering org to spec-driven delivery? That's Specdr.
Notes from the forge
What I'm learning while building with agents. On Substack.