A project-based roadmap of 16 weekly missions — build real apps, use AI properly, deploy your work, and create job-ready proof. AI from Day 1, as a tool, never a crutch.
Most courses hand you videos and a certificate. We hand you 16 real missions, a deadline, and a team that won't let you coast. You finish able to take a vague problem, design it, build it, debug it, ship it, and explain every decision — using AI the way working engineers actually do.
Not a prompt-monkey. Not a tutorial-follower who freezes the moment the video stops. The goal is a developer who can sit in front of an unfamiliar problem and think, build, and ship — with AI as a force-multiplier, not a replacement for understanding. That's what makes you hireable in an AI-native world.
No passive lectures. Each of the 16 missions gives you an outcome, a deadline, and a pass gate. You learn exactly what you need to ship it — then you ship it: a GitHub repo, a live deployed link, and a Loom walkthrough explaining what you built and why. The rule that runs the whole programme: if you can't explain it, you can't submit it.
Every mission has an explicit AI policy that escalates as you grow — from AI as Tutor, to Reviewer, to directing an AI coding Operator. AI explains, generates, and reviews; you decide, verify, and own the result. We grade your understanding, not your output — which is why proof beats promises and nobody hides behind a generated answer.
Every mission runs the same engine — Learn → Build → Deploy → Explain → Review → Improve — so the habit compounds week after week. The phases take you from Foundations to Real Apps to AI Products to a Capstone, with optional Honours missions for stronger candidates who want to go further.
Every mission kicks off live on Monday evening (GMT+8) — deliberately after work and after class, so people with real jobs and real lives can actually show up. From there, daily Buildrooms and morning Unblock Hours keep the week moving. We don't clock your hours — but real skill takes real reps, so expect to invest 25+ hours a week actually building.
Every learner carries a live status — it's how we keep the programme honest and keep you moving when things get hard. We measure shipped, explained missions, never attendance. The moment momentum slips, a real human reaches out — not a reminder email.
You don't watch videos. You deploy a real app, repo, or product feature — and explain how it works.
You don't pass a week until you can explain the decisions, edge cases, and what AI helped with vs. what you verified.
AI explains, generates, and reviews. You decide, verify, and own the result. Every mission has an AI usage log.
Every mission ends with a GitHub repo, deployed link, Loom, and public LinkedIn + X log entry.
Every mission has an explicit AI policy. AI is treated as a tool you direct, not a substitute for understanding. The policy evolves as missions progress — and there's one mandatory NO-AI gate to prove independent debugging.
AI explains concepts, debugs errors, suggests edge cases. You write code, decide, and verify. AI does NOT generate work you can't explain.
AI reviews architecture, suggests refactors, finds root causes. You accept or reject every suggestion and verify behaviour yourself.
You direct an AI coding agent (Codex) through a full ship cycle — spec, plan, diff review, approve, deploy. You approve plans, reject bad diffs, and own rollbacks.
AI is the feature itself. You design prompts, protect API keys, handle limitations, write golden datasets and run evals on output quality.
AI drafts PRs, case studies, outreach, resume bullets. You verify every claim is truthful. Every mission has an AI usage log — transparency is the standard.
NO AI during the 4-hour Debugging Gauntlet. You debug from docs, DevTools and code reading only. Proof you can think without the autocomplete.
Most of the stack runs on free tiers — GitHub, Vercel (Hobby), and Supabase (Free) are enough for the whole programme. AI tooling is the exception:
You are responsible for your own accounts, credits and tokens — including any paid subscription or API usage beyond free trials and free tiers. Always set a billing cap, and never commit an API key to GitHub.
A mission a week, built live with your cohort. We don't track your hours — we care whether you ship the mission and can explain it. The Programme is intensive, but the structure is yours.
The daily anchor — one hour, live with mentors and your cohort. Build alongside others, ship together, debug together. Skip a session and someone will text you the same day.
Morning drop-in if you're stuck. 1-on-1 with a mentor to break through whatever's blocking your mission before the day starts.
GitHub repo + deployed link + a Loom walkthrough explaining what you built and why. That's how we test understanding — not attendance.
Mentors and peers in the same channels you are. Async help when you don't need to schedule a call.
An entire team is in your corner — from the first line of code to a signed offer.
Deliver the curriculum. Deep subject-matter expertise, and genuinely there to teach and mentor.
Help you troubleshoot code, understand concepts, and finish your missions when you're stuck.
Maintain hiring-partner relationships and help place graduates into the right roles after the Programme.
Guide your job search — resume, interview prep, and navigating offers.
Runs the operations end to end, from scheduling to smooth weekly delivery.
Focused on retention and momentum — academic support and helping you over the hurdles.
Every learner has a status. It's how we keep the Programme honest — and how we keep you moving when things get hard. We don't measure your time; we measure shipped, explained missions.
Shipping missions on time, submitting Loom walkthroughs, asking good questions. Keep building.
Submission late or thin. A mentor reaches out — short call, recalibrate, no judgment.
We pause and have a real conversation about pace, fit, and the path forward.
You learn what you need to ship each mission — from Sigmo lessons, your mentor, the cohort, Discord, and AI.
Our learning platform — ~135 micro-lessons across 13 modules. Pull exactly what a mission needs, when you need it. A guided path, not an endless course library.
Daily cohort build sessions + morning Unblock Hours. Real humans, live, every weekday.
Your cohort, mentors, and alumni in one always-on space. Pair up, swap wins, get unstuck async.
GitHub, Vercel, and Loom — where you ship, deploy, and prove understanding on every mission.
OpenAI Codex for agentic coding and the OpenAI API for product features — used responsibly, with an explicit policy per mission.
Portfolio, case studies, mock interviews, hiring-partner intros, and a 30-day job-search plan — built into the final missions.
This roadmap is the curriculum. The Programme wraps it in live Buildrooms, a team in your corner, and a job-readiness engine.
12 core missions + 4 optional honours, grouped by phase. Every mission follows the same loop. Tap any card for the full brief.
By the end of the 90 days, the goal is proof you can build, debug, deploy, explain and improve real software. Judged by one question: “If an employer asked why they should trust this beginner, what proof can you show?”
