01
Techleap · 0→1 · Built with AI

Designing and building Techleap's membership platform with AI

Techleap, the community of the Netherlands' top founders, was going commercial — and needed a membership platform. But the real assignment was a new way of building: an AI workflow where the whole team ships, no designer–developer hand-off.

Role
Freelance product designer & builder
Team
Startup-mode team of ~8; built with 2 developers
Built with
Figma + Replit (AI workflow)
02
Business context & goals

The platform was the deliverable. The new way of building was the point.

Techleap — 500–600 curated founders, government-funded and going commercial — needed paid memberships live before its flagship event, where for the first time founders needed a membership to attend. Real stakes, but a familiar kind of project. What made it different was how it had to be built.

The bet

The team was eight people wearing several hats — no in-house designers or engineers to lean on long-term. The community director's bet, one that only made sense after AI: everyone on the team should be able to ship features themselves, without waiting on a designer or a developer hand-off. My job was to build the platform and — more importantly — make that way of working real.

I was a freelancer on a temporary contract. If the platform only worked with me in the room, I'd have failed — it had to outlast me.

03
Problem

Two sides to serve — and a way of working that couldn't get there.

Founders needed

One place to connect, find and join events, and see who's in the community.

Techleap needed

To charge memberships, collect the right founder data to curate better, promote its in-person products, and give founders a marketplace to sell to each other.

04
Approach & rationale

A new way of working: talk to users, vibe-code, review, ship.

The first version was built the traditional way — me as the designer plus two developers. Then AI changed the question. We wanted to move fast, so the goal became to vibe-code the platform itself, with a setup safe enough to actually ship. The loop we set up:

01Talk to users, define featuresMe
02Vibe-code in Replit, design system connectedMe + AI
03Open a PR on GitHubMe
04Claude reviews the PRAI
05Engineer does the final checkDev
06DeployShip
Build, not buy
An off-the-shelf tool like Luma couldn't give Techleap the control it needs — invite-only tiers, hand-picked members and events — so the platform was built around that curation rather than bent to fit a generic product.
A sensible trade-off
With ~500–600 curated members, we could accept a little less code purity in exchange for speed. At that scale, shipping fast mattered more than perfect engineering.
05
The design system

One source of truth — for the platform, and everything around it.

Built from Techleap's brand guidelines and tightened, then connected into Replit and used by Claude — it kept every step of the loop on-brand without a designer in the room. It became the backbone of Techleap's communication: the platform, the new website, and their social and printed content.

The Techleap design system documented — gradients, neutral scale, semantic color tokens and components
The design system, documented and wired into the AI workflow.
06
Feature spotlight

Circles: the intimate space founders kept asking for.

The gap: founders told me Techleap was great at big events — but had no intimate space to go deep.

The answer: Circles — 7–8 founders at the same stage and vertical, meeting quarterly. First one: B2C founders at growth stage.

My part: designed the format from founder conversations, then ran it with a colleague — facilitating sessions and feeding what I learned back into the product.

A Circle on the platform — B2C Growth: what to expect, who's in the circle, join the waiting list
The B2C | Growth Circle on the platform — what to expect, who's in, and the waiting list.
07
User flow

A Circle, end to end.

1Founder

Browse and apply

Circles are listed on the platform; founders apply by answering 5 key questions.

2Techleap

Review and approve

Applications reviewed, members approved — keeping each Circle curated.

3Platform

Activate — space created

The platform auto-creates a shared space: sessions, challenges, documents, group chat.

4Host

Prepare the challenge

The session's host writes a challenge; the platform guides them to prepare it well.

5Founder

Send it round

Everyone gets the challenge by email before the session, with time to prepare.

6The Circle

The session

The host takes the hot seat; members follow the challenge on their phones to think along.

7Facilitator

Post the takeaways

The facilitator, recording and taking notes, posts the main takeaways afterward.

B2C Circle home — members, cohort progress, next session
The Circle's space
Prepare your challenge — six questions overview
Guided preparation
Post a challenge — step 1, the core issue
Posting a challenge
08
The shipped platform

The whole platform.

Posting and charging for events, memberships and Circles, a member network and directory, and founder data — the features came from talking to founders about what they actually needed, and from what Techleap needed to run a commercial community.

09
Impact

Not one launch — an operation that can run itself.

01
550+ paying members
Shipped in time to take the first paid memberships before the big event. Today around 550 founders pay for membership and manage it through the platform.
02
6 Circles live — 21 the goal
The first launched in April 2026; the plan is 21 by the end of the year, all running on the platform end to end.
03
A team that could carry on
Techleap can charge, curate, run events, and keep shipping on a process and design system that don't depend on me.
10
Reflection

What I learned: relearning how to work.

Adapting to a new workflow, on the job

I had to relearn how I work — and leveraging AI tools meant shipping roughly ten times faster than I could have before. Feedback on the platform has been great, and the process has let the whole team produce and experiment in a new way.

Velocity as a value

Velocity matters to Techleap — the point was always to fail fast and try new things. Matching the engineering to the scale (a few hundred curated users) made that possible: speed beat purity.

That was the point: not one launch, but an operation that can run itself.