Eleven famous companies hiring AI & LLM engineers in Singapore right now — ranked by how close the work sits to an actual frontier model.
DeepMind chose Singapore as the headquarters of its flagship applied-AI team for all of Asia-Pacific. This is the rare chance to ship Gemini-class models into products without leaving the island — research-adjacent work, production stakes.
The deepest pool of LLM engineering work in Singapore by sheer volume. The Seed team builds ByteDance's foundation models, and the local search team ships LLM-powered AI Search, agents and RAG systems into products used by billions.
OpenAI planted its first applied-AI lab outside the US in Singapore, backed by a US$235M commitment — 37 Singapore-tagged openings verified live. Most are commercial, but a genuine engineering core is hiring now, including a quietly fascinating consumer-hardware track.
The maker of Claude began advertising its first Singapore-based roles this month — finance, support, contracts and a research economist (S$307k–331k). No engineering postings yet, which is exactly why it belongs on a watchlist: first eng roles tend to follow within quarters.
Singapore's own research lab, inside NYSE-listed Sea. SAIL publishes serious foundation-model and reasoning research and runs an Industrial PhD Program with local universities — full tuition plus salary. Smaller than the giants, closer to the research.
Qwen is among the most-used open-weight model families in the world, and Alibaba co-built Qwen-Sea-Lion-v4 with AI Singapore. But the engineering happens in Hangzhou and Beijing: a full sweep of both official career portals (1,341 live postings) found just three Singapore roles — and none of them is AI engineering. Watchlist only.
The quiet surprise of the sweep: Tencent runs genuine foundation-model research out of its Singapore office, not just sales. A live Hunyuan reinforcement-learning role sits at CapitaSky alongside Youtu Lab's multimodal-LLM team and a coding-LLM research line — this is frontier engineering on the island, not a deployment outpost.
Everyone else's models run on NVIDIA's silicon, and its Suntec Tower team is how that lands in Asia. The Singapore roles are GenAI solutions architecture, developer relations and data-centre networking — customer-facing AI engineering at the centre of the boom, even though Nemotron itself is built stateside.
Apple keeps its foundation-model R&D in Cupertino, but two genuine ML engineering seats are open in Singapore: a GenAI & ML role doing RAG, fine-tuning and agentic apps, and a Siri language engineer working on-device NLP. Quiet, well-paid applied AI inside the most valuable company on earth.
Singapore's own decacorn runs one of the region's largest production-ML operations — pricing, fulfilment, fraud and an internal GrabGPT layer — all engineered out of the One-North HQ. Not a foundation-model lab, but if you want LLMs and ML shipping to millions of daily users from a Singapore desk, the seats are here.
The national programme, hosted at NUS, that builds SEA-LION — the Southeast-Asian LLM that Alibaba co-trained into Qwen-Sea-Lion-v4. Smaller and mission-driven, but it is the one place on this list where the model being built is about the region, and the engineering genuinely happens here.
If you want famous and building the LLM itself, the strongest plays today are Google DeepMind — the only Western frontier lab running core applied-model engineering out of Singapore — and ByteDance Seed, which simply has more open LLM engineering seats here than everyone else combined. The expanded sweep adds a third name to that tier: Tencent quietly runs Hunyuan and Youtu-Lab research out of CapitaSky — genuine foundation-model work, not a sales outpost.
OpenAI is further along than the headlines suggest: of 37 verified Singapore openings, ten are technical. Anthropic just arrived with four non-engineering roles — being early to a lab's first office is historically a very good trade. Apple keeps two applied-GenAI seats open, and NVIDIA offers nine field/solutions roles at the centre of the hardware boom — neither builds its model here, but both put you next to the work.
For homegrown plays, AI Singapore is the one place on the list building a model about the region (SEA-LION), and Grab ships production ML to millions from One-North. Several famous labs — Meta, Mistral, Cohere, xAI, Hugging Face — turned up zero Singapore AI roles in the sweep, and were left off.
Practical move: apply to DeepMind, ByteDance and Tencent now if you want to build the model; set alerts on OpenAI, Anthropic and Apple; and keep Sea AI Lab, AI Singapore and Grab in play depending on whether you lean research, mission, or scale.