Skip to content
Singapore · July–August 2026 · Application-based

Build with AI. Judge with taste.

A small in-person cohort for product practitioners who can now build with AI and want the design and engineering judgment to know whether the work is any good.

Across a five-week live arc, plus pre-work, you will build, use, improve, and defend two permanent artifacts: your personal design system and your personal AI agent.

Built for practitioners already shipping products and owning decisions about what gets built.

10–14 buildersA small Singapore room, calibrated by application.
60–70 hoursFive-week live arc, plus 8–12 hours of pre-work.
Two artifactsYour design system and your AI agent, both in daily use.
One standardCan you justify every consequential product decision?
What you will walk away with

Two things you keep. Both have to hold up.

AI can generate artifacts in minutes. This course teaches the judgment to decide whether they are coherent, durable, useful, and actually yours.

01

Your personal design system

Tokens, type, color, spacing, core components, interaction principles, and a vocabulary for critique. Created under a practicing designer's instruction and reused in everything you build after.

  • Tokens, type, color, spacing, and components
  • Interaction patterns and interface principles
  • A reusable foundation for future AI-assisted builds
02

Your personal AI agent

A small, deployed AI-powered workflow or interface built around your own work. You are the user, so the usage data is real from day one.

  • Scoped to a real recurring workflow
  • Deployed, instrumented, and used daily
  • Improved through usage data and peer testing
03

The fundamentals behind both

Design hierarchy, contrast, density, patterns, observability, testing, cost, security, drift, and an operator-grade mental model of how LLMs work.

  • Practice with real generated artifacts
  • Structured critique from senior practitioners
  • A standard you can reuse after the cohort
Defense Day

The agent is the vehicle. The judgment is the point.

Defense Day is a live critique panel where you defend your design and engineering decisions. You will not just demo the thing. You will defend the decisions behind it: scope, interface, architecture, cost, security, observability, drift, and what you would change next.

01Design coherence

Do the system choices create a usable whole?

02Interface clarity

Is the output clear, or merely polished?

03Architecture

Will the agent survive real use beyond the first demo?

04Cost

Can the workflow scale without surprise spend?

05Security

Is it safe to connect to your own data?

06Observability and drift

Can you see what changed and correct it?

Who this is for

For practitioners who currently ship products.

Most of you will be PMs, product leaders, founders, operators, designers, or engineers doing product-shaped work. Your title is not the filter. Your work is.

You will fit if you…

  • Can ship, but want to know if what you shipped is production-grade
  • Can generate ten UI variations, but want sharper taste for choosing
  • Have shipped a build where the landing page and the next screen feel like two different products — and want one system so it never happens
  • Want to move past local scripts to always-on agents that keep running when you're away from the keyboard
  • Want to keep knowing your system behaves as intended, with evaluation built in
  • Want a design system and AI agent that are actually yours

What this is not

  • Not a coding bootcamp. AI does the typing.
  • Not a Figma course. Fundamentals for judgment, not tool training.
  • Not prompt-engineering training or an AI PM course.
  • Not for breaking into product or passive workshop attendance.
How it works

A five-week live arc, plus pre-work.

The course moves from fundamentals to artifacts, then from artifacts to lived use, iteration, peer testing, and final defense.

0Pre-work
8–12 hours before Day 1

Design literacy, engineering literacy, and LLM mental models

Complete guided pre-work and a diagnostic so the live days can start at full speed.

1Day 1
Saturday in July · 9:30am–5:30pm

Scope your agent and build your design system

Pressure-test the LLM mental model, defend your agent scope, then build the first version of your personal design system with guidance from a practicing design leader.

2Day 2
Saturday in July · 9:30am–5:30pm

Engineering fundamentals and agent build

Learn architecture, testing, observability, cost, security, and drift. Then build your agent using your own design system.

3Build
Three async weeks · 5–10 hours per week

Deploy, instrument, use, and iterate

Live with the agent daily. Includes cohort check-ins, specialist office hours, structured peer testing, iteration prompts, and usage-data review.

4Defense
Saturday in August · 10:00am–6:00pm · dinner follows

Defend every consequential choice

Present both artifacts and defend scope, design coherence, architecture, cost, security, observability, drift, and what you would change next.

Who's in the room

Real practitioners teach the fundamentals, then judge the application.

The course's differentiation is that the people teaching design and engineering judgment return on Defense Day to press on how you used it. Engineering faculty will be announced once confirmed.

Engineering faculty

To be announced

Teaches engineering judgment: patterns, architecture, observability, testing, cost, security, and drift prevention. Joins the Defense Day panel once confirmed.

Keith Oh

Head of Product Design, Carousell

Teaches design judgment: hierarchy, contrast, density, interaction patterns, design systems thinking, and the vocabulary of critique. Joins the Defense Day panel.

Deepika Murthy

Lead instructor

Nineteen years in product. Most recently Head of Product, Marketplace & Fulfillment Platforms at Gojek. Co-founder of Women in Product Singapore.

Support & community

You're not doing this alone — during the cohort or after.

The agent and design system are what you build. The room is what you keep.

WhatsApp support, all week

A cohort WhatsApp group with support from the team through the week — not only on the live days.

A room you keep

You leave with a small group of practitioners navigating the same shift, and an open invite to future cohorts' defense days.

Support beyond the five weeks

The cohort doesn't end on Defense Day. The community is the part that keeps compounding.

Timing, pricing, and application

The commitment is visible before you apply.

Cohort 1 is application-based and closes when the room fills. The fit call checks whether you are currently shipping products and own decisions about what gets built.

Founding cohort

Pricing

SGD 499

Cohort 1 is priced below future cohorts in exchange for being the first group through the course.

Dates

July–August 2026

Day 1, Day 2, and Defense Day are three Saturdays in July–August. Central Singapore venue shared with confirmed cohort members.

Refunds

Refunds and deferrals

Full refund until 14 days before Day 1. After that, one free deferral to Cohort 2. No refunds after the course begins.

Apply

Ready to apply for Cohort 1?

The application takes about five minutes. We review every one individually, then have a short fit call before confirmation.

4 builders have applied

Already applied? Check your status

Applications close when the cohort fills.

Spread the word

Know someone building in the merged role?

Cohort 1 is small and application-based. Send it to the practitioner who should be in the room.