Recruitment agency
CV screening — 80% time back
Consultants lost half their week to manual screening, with a three-day lag that cost them candidates. Built a screening pipeline that handles the first pass.
— For agencies & lean teams
Find the fix your client needs. Then ship it.
9 yearsof shipping products for agencies and lean teams. I find the leak worth fixing, write the spec engineering wants to build, and unblock people. That’s the whole job.
— Three problems I keep seeing
Too many ideas, no time to test them
You've got ten things you could build for the client. You don't know which one is real.
I pressure-test them. Usually the wrong one is obvious once someone writes the assumptions down.
AI feature that nobody uses
Your team shipped an AI feature for the client. The demo was great. Nobody came back.
Demos aren't products. We figure out the actual loop — input, output, what happens when the model's wrong, and why anyone would use it twice.
Roadmap full, velocity tanking
You're shipping. Nothing's moving. Nobody on the account wants to cut anything.
Half of it shouldn't ship. I'll tell you which half. Engineering will be happier than they're letting on.
— How it goes
Four steps. That’s the whole job.
Step 01
Figure it out
Week one. I read the code, sit with your team, find what's actually broken.
Step 02
Write the spec
One doc. What we're building, what we're cutting, how we know it shipped.
Step 03
Ship it
Engineering builds. I unblock, push back, rewrite when wrong.
Step 04
Hand it back
Your team owns it. I leave notes, not dependencies.
— Usually 6–12 weeks. Sometimes less.
— In practice
What that looks like in practice.
Each started as a diagnosis. The build came after I found the leak worth fixing — not the one they first complained about.
— How I work
Same way every time.
Doesn’t matter the team or the stage. It’s usually the same three problems.
01
Nobody's sure what we're building.
Discovery first.
A week figuring out what we're actually doing before anyone writes code. Skip this and you'll build it twice.
Pick the number that matters.
If the team can't name the one metric that moves, the roadmap is theatre.
02
Priorities are fighting each other.
Decisions go in a doc.
Same argument doesn't happen twice. Nobody has to remember what we agreed in April.
Async. Meetings capped.
I read code. I write updates. I push back when you're wrong. No status theatre.
03
Tickets are vague.
Specs engineers want to build from.
AC, edge cases, QA steps. Not because process — because it's faster than guessing.
AI in the workflow, not just the product.
PRDs, ticket refinement, edge cases, retros. I'll show you how. It's not magic, it just saves hours.
— How to hire me
Embedded
Be your PM
Run the whole thing.
I'm your product person on the account. Discovery through delivery. Async, meeting cap, 2-month minimum.
FOR · Your agency doesn't have a product lead. You probably need one. You don't want to hire full-time yet.
Pricing on request
Book a call →10 days
Sprint
One problem, ten days.
Pick one thing — the brief you can't shape, the AI feature that flopped, the roadmap that's full of nothing. Ten days, one output.
FOR · You're stuck on something specific. You want a deadline and fresh eyes.
Pricing on request
Book a call →Hourly
An hour at a time
No commitment.
Audits, second opinions, spec reviews, pitch prep. Sometimes an hour is enough.
FOR · You want a senior PM's take before you commit to anything bigger.
Rate on request
Book a call →
— What happens next
Tell me what you’re building.
Step 01
You email what you're stuck on.
Step 02
15-min call, no slides.
Step 03
I send a one-pager with what I'd do and a price.
Step 04
We decide if there's a fit.