AIAP

AIAP's design is highly distinctive:

• Apprenticeship: not classes, but paid work on real projects
• 9 months: month 1 is deepskilling (intensive ML fundamentals), months 2-9 are real 100E projects
• Mentorship: each apprentice cohort is paired with senior AI engineers / researchers as mentors
• Strict selection: 1,000+ applicants per year, ~60-80 accepted

Apprentice backgrounds:

• Cross-discipline transition: traditional software engineers moving into AI (largest group)
• Fresh graduates: CS / mathematics / physics backgrounds
• Returnees: AI master's / PhD graduates from overseas returning to Singapore
• Industry transition: IT staff from finance, healthcare, etc. moving into AI

After graduation, around 70% stay in Singapore's AI industry; 30% leave for overseas big tech.

The core problem AIAP solves: a severe shortage of locally trained AI engineers in Singapore.

Singapore's universities produce high-quality but small numbers of CS graduates (NUS / NTU together produce roughly a thousand per year), and only a fraction actually go into AI. Commercial AI teams (DBS, Singtel, Grab) and startups are constantly competing for talent. AIAP gives engineers from "non-CS backgrounds with strong learning ability" a fast track into AI engineering — effectively a "side door" into the local AI talent market.

Technically, AIAP's training intensity is very high — going from zero to writing production-grade AI projects in 9 months requires apprentices to work 10+ hours a day. That intensity filters out those who aren't suited; what remains are people who can genuinely deliver.

AIAP is the core lever for "Lever 2 (talent)" in Singapore's AI strategy.

In the seven-lever framework:

• Lever 2 (talent): the main pipeline for training local AI engineers
• Lever 3 (application): through 100E projects, apprentices get real enterprise experience

A take: AIAP's "apprentice + project" model is the biggest innovation in Singapore's AI talent strategy. Universities cannot do this (too academic), enterprises cannot do this (unwilling to spend 9 months slowly mentoring newcomers) — only an institution like AISG, sitting on the "national programme + commercial project bridge", can pull it off.

But AIAP also has structural challenges: high attrition — after 9 months, apprentices move into private sector roles and AISG cannot retain its own graduates; project quality depends on the enterprise side — the calibre of 100E partner projects varies significantly; scale ceiling — 60-80 graduates per year is still a drop in the bucket for Singapore's AI industry.

The key questions for AIAP in the NAIS 2.0 era: Can it scale to 200+ per year? Can retention be improved? Can it produce higher-calibre AI engineers?