Radar

Skills radar

Indonesia & Southeast Asia - focus on AI / ML / data / platform engineering. Snapshot for 2026-07.

20 skills

Skill demand

1Pythonstable98
Default language across almost every SEA AI/data posting.
Practice in Lab →
2SQLstable92
Still the highest-ROI skill for analysts and ML people who touch warehouses.
Practice in Lab →
3Technical Englishstable90
Remote-friendly SEA roles and SG/ID multinationals filter on this hard.
4PyTorchup86
Preferred over TensorFlow in new ML Engineer JD language.
Practice in Lab →
5RAG / retrieval systemsup84
Enterprise AI projects in ID/SG almost always include private document Q&A.
Practice in Lab →
6LLM APIs (OpenAI / Claude / Gemini)up82
Shipping assistants beats training foundation models for most local teams.
Practice in Lab →
7Dockerstable80
The minimum bar for “can you deploy what you built?”
8scikit-learnstable78
Tabular baselines still dominate banking/fintech use-cases.
Practice in Lab →
9pandas / data wranglingstable77
Messy SEA enterprise data makes wrangling non-optional.
Practice in Lab →
10Kubernetesup74
Shows up for platform and serious ML serving roles; overkill for many startups.
11OpenCVstable72
CV roles in manufacturing/agri still list classical vision + deep learning.
Practice in Lab →
12Airflow / orchestrationstable70
Data engineering postings lean Prefact/Airflow/Dagster language.
13Spark / big datadown68
Still in banks/telco; many startups replaced it with warehouse ELT.
14FastAPI / model APIsup67
Notebook → service is the hiring filter for applied ML engineers.
Practice in Lab →
15Vector DBs (pgvector / etc.)up66
RAG stacks need somewhere to put embeddings; pgvector is common.
16AWSstable65
Most common cloud in ID enterprise JDs; GCP strong in some AI teams.
17MLOps / experiment trackingup63
MLflow/W&B language appears once teams leave prototype stage.
18LangChain / agent toolingup61
Often listed; seniors care more about evals than the framework name.
Practice in Lab →
19AppSec / secure SDLC basicsup58
Fintech/health buyers ask about it even for AI features.
Practice in Lab →
20TensorFlow / Kerasdown52
Legacy stacks and some mobile/edge teams; fewer greenfield JDs.

Roles in focus

ML Engineerhigh
Most postings want PyTorch + serving basics, not just notebooks.
Data Scientisthigh
Still strong in banks, e-commerce, and telco - SQL + experimentation language wins.
Data Engineerhigh
Often better paid than DS for the same years of experience in Jakarta.
AI / LLM Engineerrising
RAG + eval + cost control show up more than “prompt engineer” alone.
Computer Vision Engineersteady
Manufacturing, agri, and security use-cases; OpenCV + deployment matter.
Platform / MLOpsrising
Fewer pure titles, but every serious AI team needs this skill mix.

Market signals

RAG > “we need an LLM team”
Many Jakarta/SG postings now describe retrieval, evaluation, and cost controls - not training foundation models.
Serving skills separate juniors from hires
Docker + a simple model API often matters more than another Kaggle certificate.
CV remains niche but real
OpenCV + YOLO-style skills show up in manufacturing, agri, and security - smaller market, clearer problems.
Data engineering stays expensive
SQL + pipelines + cloud warehouse work is still easier to monetize than pure modeling in many local firms.

Close the gap

Learn the skills that show up in real JDs — not just tutorial checklists.

Explore the Lab