AI-Native · Indonesian-context · Free
AI-Native engineering. Remote-ready salaries.Free.
Twelve weeks of AI-Native engineering — agents, RAG, evals, fine-tuning — learned by simulating real work inside a fictional Indonesian SOE codebase. Indonesian engineers ship from here into remote AI roles paying $150–200K/year (≈ Rp 2,4M – Rp 3,2M/yr) or into Komdigi-rate junior tracks at home.
Built by Metatech. Aligned with Komdigi DTS & Kepmen Kominfo 117/2024.
What ‘AI-Native’ means here
Built around how AI engineers actually work today.
Five things we teach. None of them are AI in isolation — the model, the product around it, the safety layer, the delivery toolkit, and the simulated operations to put it all under pressure.
AI-Native, not AI-Bolted-On
We don’t teach import openai and call it AI engineering. The curriculum mirrors how engineers actually ship AI today: agents, RAG, evals, fine-tuning, multimodal. From day one.
Full-stack, shipped via coding agents
AI Engineering isn’t just the model. It’s the full-stack product around it — frontend, backend, deploy, monitoring — built day-to-day with Cursor and Claude Code. You learn to direct a coding agent to ship features, not to type every line.
Safe, compliant, ready for real users
An AI product that’s safe to ship to Indonesian users means UU PDP compliance, content moderation, prompt-injection defenses, sensible defaults. Pandai teaches the policy + safety layer alongside the technical layer — because AI products without it shouldn’t ship.
Delivery toolkit: ship like an operator
Real engineers don’t just push code. They write the slides for the demo, draft the timeline, operate the spreadsheet, generate the image or animation for the presentation, fill in the admin doc the regulator wants. Pandai teaches all of that with AI — whatever ships fastest.
Simulated real-work experience
You ship every Friday into a fictional Indonesian SOE codebase (PT Nusantara Logistics). Five years of half-finished digital transformation. Production incidents. On-call rotations. Scope changes. Not assignments — operations.
See The Yard →Remote · Global employers
Indonesian engineers earn this much already. Remote.
Three real comp bands — two remote from Indonesia, one in Silicon Valley. USD per year, Rupiah equivalent at IDR/USD ≈ 16,000. Different currency, same engineer. The AI-Native job market hires Indonesians who can ship.
Software Engineer
FAANG company
Onsite in HQ, Silicon Valley
$600K /yr
≈ Rp 9,6M /yr
Indonesians placed at FAANG HQs — public Levels.fyi bands.
AI Engineer
APAC company
Remote from Indonesia
$180K /yr
≈ Rp 2,9M /yr
Where Pandai’s founder works.
Engineer
US company
Remote from Indonesia
$153K /yr
≈ Rp 2,4M /yr
Indonesian engineers shipping product to a global audience.
Public salary data from Levels.fyi and company filings; one band disclosed by Pandai’s founder. Indonesian-passport engineers, remote contracts. Rates vary by level, scope, and equity.
Indonesia · Komdigi 117/2024 rates
Or stay local. Junior → Middle → Senior under Kepmen 117/2024.
Five tracks. Three tiers. Pandai is the on-ramp at Junior and the lift to Middle / Senior — same Komdigi rumpun, same regulated rates. The headline is Senior AI Engineer: Rp 150jt/mo ≈ Rp 1,8M/yr.
Engineer — AI
Agents, RAG, fine-tuning. Premium track. Senior ceiling is the headline.
junior
Rp 25,6jt /mo · Rp 307,2jt /yr
middle
Rp 28,6jt /mo · Rp 343,2jt /yr
senior
Rp 150jt /mo · Rp 1,8M /yr
Engineer — Software
BE / FE / Mobile, AI-augmented day one.
junior
Rp 15jt /mo · Rp 180jt /yr
middle
Rp 20,8jt /mo · Rp 249,6jt /yr
senior
Rp 46,4jt /mo · Rp 556,8jt /yr
Engineer — Data
Data Analyst + Data Engineer.
junior
Rp 12,5jt /mo · Rp 150jt /yr
middle
Rp 28jt /mo · Rp 336jt /yr
senior
Rp 49,6jt /mo · Rp 595,2jt /yr
Engineer — Quality
AI-assisted testing. QA-AI discipline.
junior
Rp 17jt /mo · Rp 204jt /yr
middle
Rp 28jt /mo · Rp 336jt /yr
senior
Rp 35jt /mo · Rp 420jt /yr
Tribe Operation
AI-augmented PM, scrum, project admin work.
junior
Rp 14,9jt /mo · Rp 178,8jt /yr
middle
Rp 20,6jt /mo · Rp 247,2jt /yr
senior
Rp 42,3jt /mo · Rp 507,6jt /yr
Source: Kepmen Kominfo 117/2024 rate structure (rumpun × level lookup). Annual estimates assume 12× monthly without bonus. The in-person cohort places students at Junior; the free curriculum + community lift existing engineers to Middle/Senior — see cohort details →
Suggested progression
Start anywhere. The path goes M0 → M11.
Modules unlock as you ship — no paywall, no gating. The first two are the on-ramp; the rest read better in order. Pull what you need.
- Available
M0
Pre-week — setup, AI-Native habits
Curated YouTube videos + Pandai SHIP project
- Available
M1
AI-First Coding & IDE (Cursor / Claude Code)
Curated YouTube videos + Pandai SHIP project
- Up next
M2
LLM Fundamentals — tokens, cost, eval
Curated YouTube videos + Pandai SHIP project
- Suggested after M2
M3
RAG — embeddings, retrieval, harnesses
Curated YouTube videos + Pandai SHIP project
- Suggested after M3
M4
Agents I — tool use, planning loops
Curated YouTube videos + Pandai SHIP project
The 12-week arc
Three phases. Twelve weeks. One arc from prompt to production.
Vibe coding → engineering depth → track specialization. Each module pairs curated videos with a Pandai SHIP project.
AI Vibe Coding
Claude Code → LLM fundamentals → Agents + MCP → Practical shipping
Ship something useful with Claude. By Week 4 you’ve built an agent that does real work — video, slides, Excel, Indonesian forms, or website registration.
Engineering Depth
Bella case study → Production deploy → Pick your track
From toy to product. End-to-end LLM apps with auth, evals, observability. Real deploys to a fictional Indonesian SOE.
Track Specialization
AI Engineer · Software · Data · Quality · Tech Ops
Pick your specialty. AI Engineer ends in self-improving closed-loop systems (autoresearch), not fine-tuning. Capstone for a real partner.
SHIP projects
Ship every Friday. Public GitHub.Real users.
Each module ends with a SHIP project. You build it, you push it, you record a Loom walking through what you built and what you got wrong. Not assignments. Real shipped artifacts on a public repo.
Operational simulator
Coding bootcamps train coders. Pandai trains operators.
The Yard — PT Nusantara Logistics
A fictional Indonesian SOE you can work inside, starting today. Production incidents, on-call rotations, scope changes, regulator deadlines — the operational reality real engineers face after they get the offer.
What makes Pandai’s curriculum defensible
Four moats. None of them are easy to copy.
AI-Native discipline
Built on eval, harness, and agent discipline — the actual moat. Agents, RAG, evals, fine-tuning, multimodal — from day one. Not Python lessons with an AI sticker on top.
Metatech engineering DNA
FDE skills taught only by people who’ve shipped FDE work. Karya1 to BGN. Bella, Lovina, Jalinworks. The instructors built the things they’re teaching.
The Yard — simulated real work
PT Nusantara Logistics: a fictional Indonesian SOE codebase with five years of half-finished digital transformation. Production incidents, on-call rotations, scope changes — not assignments.
Bahasa-native, Indonesian-context
Indonesian tokenization, fine-tuning, low-resource RAG. UU PDP / OJK / BSSN compliance baked in from day one. KTP fields, Rupiah formatting, Bahasa-native UX throughout.