Home / Courses / Engineers × Founders Bootcamp
Practitioner Course — 3-Day Bootcamp

Engineers ×
Founders

A practitioner course for engineers partnering with founders. Each engineer is paired with one founder on Day 1 — together they take a vision from idea to a production-ready MVP in 72 hours, then keep building as partners. Built and taught by operators who ship AI products at scale.

⚡
Duration
3 days
👥
Pairing
1 engineer : 1 founder
🎯
Deliverable
Live, deployed MVP
🔥
Format
Teach · Build · Demo
📅
First cohort starts
June 5, 2026
Join the Waitlist See the 3 Days
Program Philosophy
Engineers don't take
tickets. They translate vision.
The best AI engineers shape what gets built, not just how. They turn a founder's idea into the fastest path to a working product — with AI as the primary building block, not an add-on.
Engineers don't just build what founders describe. They translate vision into the fastest path to a working product — using AI as the primary building block, not an add-on.
— PROGRAM PRINCIPLE
Sent 3 Days Before
Pre-work matters.
Day 1 starts at 9 AM with the environment ready and the problem already in your head. Pre-work is how we get there.
📝

Read

Founder vision doc — a 1-pager submitted by each founder ahead of time, so engineers walk in already chewing on the problem.

🔧

Setup

Dev environment, API keys (OpenAI / Anthropic / Gemini), GitHub, and Vercel / Railway / Render accounts — ready to deploy.

🎬

Watch

20-minute primer on LLM APIs, embeddings, and RAG basics so the morning teaching block lands faster.

📝

Complete

"AI Capability Menu" — a reference sheet of what AI can do out of the box vs. what needs custom work.

The Curriculum
Three days.
Idea to production.
Each day pairs a morning teaching block with an afternoon build sprint and an evening demo + retro. Every day ships a tangible deliverable.
Day 1 of 3

Understand, Align, Architect

Theme: Don't build yet.
Get obsessed with the problem before you write a line of code.
Morning — Teach
9:00 AM – 12:00 PM
  • AI-First Product Thinking: the product is the model's behavior, not just the UI
  • The AI-First Stack: Input → Context → Model → Output → Action
  • Workshop: mapping AI touchpoints in 10 real products
  • The 3 AI product patterns: Generator, Classifier, Agent
Midday — Pair
12:00 PM – 1:00 PM
  • Founder-Engineer pairing & structured discovery
  • The 5 Questions Framework (say vs. mean, the one output, what's "good enough")
  • "What would feel like magic?" exercise
  • Working lunch: founder + engineer eat together
Afternoon — Build
1:00 PM – 5:30 PM
  • The AI MVP Stack: Next.js / FastAPI, LLM API, vector DB, simple auth
  • "Prompt is the product" — writing system prompts that encode the vision
  • Live demo: build a working AI endpoint in 30 minutes
  • Build sprint + 4:30 check-in: kill scope, find the Core Loop
Day 1 Deliverable
GitHub repo with defined tech stack, system prompt v1, data schema sketch, and core loop diagram (user action → AI response → user value). Evening: 5-min core-loop demo + peer feedback. Overnight: stress-test the prompt with 10 edge cases.
Day 2 of 3

Build the Core Loop

Theme: Make it real, then make it reliable.
The AI does the right thing 80% of the time? That's a product. 60%? That's a demo. Know which one you have.
Morning — Teach
9:00 AM – 12:00 PM
  • Retro: what broke overnight? Share edge cases
  • Context engineering: system prompts, few-shot, dynamic context
  • RAG vs. fine-tuning vs. just better prompting (cost / time tradeoffs)
  • Graceful failure: fallbacks, confidence thresholds, human-in-the-loop
  • Streaming & latency perception — making AI feel fast
  • Live debug: take a broken AI feature and fix it in front of the room
Build Sprint 1
12:00 PM – 3:00 PM
  • Build only the core loop — one input, one AI step, one output
  • Founder available but not hovering — answers questions, doesn't direct
  • 12:30 — is data flowing into the model?
  • 1:30 — is the output structured and useful?
  • 2:30 — has the founder tested it and reacted?
Afternoon — Production-Aware
3:00 PM – 5:30 PM
  • "Demo works" vs. "product works": rate limits, error states, empty states, loading states
  • Auth, basic security, and not exposing your API key
  • Build Sprint 2: error handling, loading states, one "wow" UI moment
  • Pair Review: founder uses it solo for 10 min, engineer watches silently
Day 2 Deliverable
A working, deployed link the founder can share with one real potential user tonight. Optional User Testing Night (6–8 PM): founders text 3 people in their network; engineers monitor logs & error rates live; group reconvenes to share what real users actually did.
Day 3 of 3

Ship It and Own It

Theme: Production is a mindset.
From MVP to something real people use. Then a clean handoff so the founder can keep going.
Morning — Teach
9:00 AM – 11:00 AM
  • What "production" means for AI: observability, logging inputs/outputs, cost tracking
  • Evaluating AI quality without a test suite — building a simple eval loop
  • The engineer's ongoing role: prompt versioning, model updates, drift
  • How to hand off an AI product — what docs the founder actually needs
  • Workshop: build the 1-page "AI Brain Doc"
Final Build Sprint
11:00 AM – 2:00 PM
  • Functionality: core loop end-to-end; error states handled; edge cases addressed
  • Production basics: stable deploy URL, env vars secured, rate limits, cost estimate
  • AI reliability: prompt version-controlled; 10 test cases passing; logs in place
  • Handoff: AI Brain Doc + README; founder can redeploy without the engineer
Afternoon — Demo
2:30 PM – 6:00 PM
  • Final presentations — 10 min per pair to a panel of facilitators + peers
  • Format: Problem & Vision (founder) → AI demo (engineer) → Core prompt & why
  • Panel feedback: is the AI doing real, specific work? Would a user pay for this?
  • Closing: the engineer's role in AI-first companies — product partner, not ticket-taker
Day 3 Deliverable
Production-ready MVP on a stable URL, AI Brain Doc, README, cost estimate at 100 users, and a clean handoff. Founder can redeploy and iterate without the engineer. Engineer leaves with a partnership-ready relationship and a portfolio piece.
Daily Rhythm
Teach. Build. Demo.
Repeat for 3 days.
Every day follows the same shape so engineers and founders find their cadence by the end of Day 1.
1

Morning Teach

3-hour teaching block delivered by practitioners who ship AI products. Frameworks, patterns, and live coding demos — not slides.

➔
2

Afternoon Build

Build sprint with the founder. Facilitator checkpoints throughout. Each pair has to produce something real — not a slide, a working artifact.

➔
3

Evening Demo + Retro

Each pair presents to peers. Engineers get peer feedback (not founder feedback). Overnight tasks set the table for the next morning.

How Engineers Are Assessed
A rubric, not a grade.
Engineers are assessed across four dimensions. The goal is to know exactly where you are — and what to push next.
Skill Beginner Proficient Advanced
AI Design Can call an API Writes effective system prompts Designs multi-step AI pipelines
Speed Builds features in hours Builds core loop in < 2 hours Working prototype in 30 minutes
Founder Comms Answers questions Proactively shapes scope Protects founder from over-building
Production Thinking Gets demo working Handles errors and edge cases Instruments, monitors, plans for scale
Recommended Stack
Boring tools.
Fast outcomes.
We default to a stack that gets a deployed AI product in front of real users on Day 2 — not Day 30.
Frontend

Next.js or HTML+JS

Fast to deploy, easy to share. Frontend should never be the bottleneck on Day 2.

Backend

FastAPI or Next.js API

Quick AI endpoint setup. One-file backends are fine for the MVP.

LLM

GPT-4o / Claude

Best instruction-following on the market. Pick one for the workshop, not both.

Vector DB

Supabase pgvector / Pinecone

Free tier, fast setup. Only added if the use case actually needs retrieval.

Auth

Clerk or NextAuth

Under 30 minutes to implement. Skip custom auth for the MVP.

Deploy

Vercel + Railway

One-click deploys. The MVP is live before lunch on Day 2.

The Engineer's Role in AI-First Companies
Five takeaways
that outlast the bootcamp.
Day 3's closing session distills the engineer's mindset for partnering with founders long after the workshop ends.
Principle Why It Matters
Product partner, not ticket-taker The best AI engineers shape what gets built, not just how.
Prompts are code Version them, review them, improve them like any other production artifact.
Speed is a feature The faster you can test an AI idea, the better the product gets.
Founders need translators Your job is to say "here's what AI can actually do" with confidence.
Ship before it's perfect An AI that works 75% and is live beats one that works 95% and isn't.
Cohort applications open

72 hours.
One real product.

Sign up for the newsletter to follow along, or join the waitlist to lock in your seat. Cohorts are kept small — one engineer paired with one founder. Every enrollment funds a student scholarship.

Join the Waitlist →

No spam. Just a heads-up when the next cohort opens.

← Back to all courses