Fractional CTO & platform builder
Marketplaces, AI systems and review platforms. Strategy through shipped code.
Most software slows down because strategy, design and engineering are split across teams. I combine them into one execution layer.
Platform work
TrustRadius
Atlassian
U Miami
Finder
Nasoni
Free field guide
Six decisions about your entity model that are cheap today, structural tomorrow, and nearly unfixable once reviews are attached to them.
Get the free guide
Available for new engagements
Discovery is fully credited toward Build.
Bryce is that rare combination of talent and experience. Time and time again others would be mired in a common perspective when Bryce would find a way to cut through to the essence. I'd hire Bryce again in a heartbeat.
Austin, Texas · Central Time · —
Featured work
The peer review and planning platform for foodservice equipment.
Chefs plan a kitchen, price it to the dollar, scan equipment serial plates with the camera to identify it, and buy on peer reviews from other operators. I built the review-to-SKU enrichment pipeline and the reputation-weighted trust model underneath it.
Six months from kickoff to both app stores.
Own your algorithm.
Every platform builds an algorithm about you. AirIntent builds one for you: what you do across Claude, ChatGPT, your desktop and your files becomes a portable intent graph you own. Multi-tenant workspaces, agent teams, a discover surface, and a Linux desktop launcher, built alone from strategy through production.
Three surfaces: web, desktop launcher, browser extension.
Your whole care team, one plan of care.
Chronic conditions rarely have one answer; they span providers. MedClinx is a two-sided marketplace connecting patients with cash-pay specialty and alternative-medicine clinics, building a plan that spans all of them and tracking adherence against it. It carries clinical-trial methodology from earlier work with the University of Miami and Nasoni into consumer care.
Launching in Austin. Open for early access.
Vision reels
ChefWorthy
Top-rated kitchen equipment, ranked by the people who use it.
Tap the speaker to unmute · auto-advances in 10s
Nasoni
AI-guided medication assistance, built into the faucet.
Tap the speaker to unmute · auto-advances in 10s
Legible AI
A smarter way to track legislation.
Tap the speaker to unmute · auto-advances in 10s
How I work
We work out which sides the platform actually has, what each one gets, where liquidity comes from, and how reputation and incentives hold it together. You get an opportunity map, an architecture spike, north-star flows, and a sequenced build plan with the tradeoffs made explicit.
Fully credited toward Build if we continue.
I build it. Product, design, and engineering in one pass, shipping to production and instrumented. One operator, not a team, so nothing gets lost in a handoff between the person who set the strategy and the person writing the migrations.
Fractional CTO work once it is live: launch, measurement, iteration, and the technical decisions that come with growth.
What made Bryce a standout contractor was his weekly updates. He'd share a weekly update showcasing data insights, hypotheses and progress towards milestones. This was a great way to bring senior stakeholders across the journey.
Inside discovery
A multi-sided platform has to work for everyone at once, and the sides want different things. The shortcut that delights a buyer is the one that costs a vendor a sale. The review that's most useful to read is the most expensive to collect. None of that survives contact with a requirements document, because a list of features has nowhere to put a tradeoff.
So discovery is drawn, not written. Every side gets a lane. Every phase gets a panel: what they see, what they do, what it costs them, and how they feel about it in their own words. The disagreements show up in week two, on a wall, while they're still cheap to settle — instead of in month four, encoded in a schema.
Definition and ideation
Two ninety-minute days. Goals and challenges, input from people who actually buy the equipment, then a teardown of how every comparable product handles each phase — for buyers and vendors separately.
The storyboard
Buyer and vendor lanes across Discover, Comparison, Evaluation, Decision and Maturation. Each cell carries the screen, the pain, the action and the sentiment, quoted rather than paraphrased.
By the end of it you know what the product has to do. More usefully, you know what it has to model — which sides exist, what an entity is, and what a review is actually attached to. That's the decision that gets expensive to reverse later, and it's why discovery earns its two weeks.
Why one operator
Normally each of those steps belongs to a different person, and the context has to survive a handoff to reach the next one. Strategy becomes a brief. The brief becomes a design. The design becomes tickets. The tickets become code, several translations away from the decision that started it.
I own the whole column. Nothing is translated, because nothing changes hands. That is the entire reason one operator can outrun five specialists on work like this.
Ten steps used to need ten people. I run them with AI agents in the loop, the same kind of systems I build for clients, which is what makes one operator viable rather than heroic. The leverage is in the tooling. The judgment about what to build and what to cut is not.
One operator. All ten.
Stack
No deck. No sales process. Just product.
Bryce is a true professional with empathy for the customer. He is a wizard at user research and his designs reflect the unique insights he always manages to find.