From Polytechnic to Y Combinator: Singapore's Collar AI Beats the Odds
Source: The Straits Times
Every few months, Singapore celebrates another homegrown founder making it in Silicon Valley — but almost always, the story starts at NUS or NTU, or after a stint at a big tech firm. The Y Combinator acceptance of Collar AI breaks that mould: both founders are polytechnic students and graduates, from...
Every few months, Singapore celebrates another homegrown founder making it in Silicon Valley — but almost always, the story starts at NUS or NTU, or after a stint at a big tech firm. The Y Combinator acceptance of Collar AI breaks that mould: both founders are polytechnic students and graduates, from neighbourhood schools and middle-class families, who got into the world's most selective start-up accelerator on their second try. It is the kind of signal that says more about where Singapore's AI talent is actually coming from than any ranking of universities.
Collar AI, founded by Singapore Polytechnic student Jamie Yau, 22, and Ngee Ann Polytechnic graduate Kazel Koh, 21, was accepted into Y Combinator in March 2026 — an accelerator that receives around 10,000 applications a quarter and admits barely 1 per cent, a bar Yau says makes them "two of the youngest Singaporeans ever" to get in. The acceptance unlocks US$500,000 (S$641,000) in seed capital, three months of mentorship, and investor attention that would normally take years to accumulate. The pair met in 2022 through the National Young Leaders Fellowship, each having already launched and shut down or exited a first start-up.
Collar started life as a defence-tech project built for Singapore's Defence Science and Technology Agency (DSTA), helping with due diligence in procurement contracts. But the duo realised defence tech wouldn't scale — security concerns would block selling to other militaries — and pivoted to finance, a natural fit given Koh's time at GIC and SoftBank. Today Collar is an AI deal agent for private-market investors: it scouts deals across sources like PitchBook and Bloomberg, runs due diligence over large document stacks, drafts investment memos, builds financial models, and maintains a "company brain" that learns from each deal. It charges US$1,000 a month after a free trial, and its prospective clients include global investment firms, venture funds, private equity houses and family offices.
The pivot from government procurement to private-market dealmaking is itself instructive. Singapore's AI start-ups increasingly find their first customer in the public sector — DSTA, GovTech, the hospitals — but the ones that scale learn early how to translate that credibility into commercial markets. Collar's plan to split its team of up to six between Singapore and San Francisco also reflects a growing pattern: build in Singapore, raise and sell in the US. Koh says YC told them they were chosen for their personalities, coachability and technical credibility — traits that have nothing to do with which university they attended.
Why it matters for Singapore: The Collar story matters beyond one start-up because it widens the definition of who gets to build AI companies in Singapore. Yau and Koh came through the polytechnic route that most of Singapore's workforce actually takes, not the elite-university path that dominates start-up coverage — and Yau is explicit that her motivation is proving success doesn't require a well-to-do family or a prestigious school. With Singapore pushing AI literacy through every layer of education, from compulsory teacher training to polytechnic AI courses, Collar is a concrete proof point that the pipeline works. Watch whether more poly-bred founders follow them through YC's doors in the coming batches.
