Engineering
Founding AI Engineer
Own the AI layer behind four ventures — orchestration, evals, prompt and data pipelines. Second engineer in.
The work
EVAA operates four AI-native ventures from Singapore: MetaKitchen, Omnis, Virtual Promoter, OpenRift. Each ships its own product surface. Underneath, they share an agent stack we are building from scratch.
You would own that stack. Model selection, evals that catch real regressions, retrieval and tool-use plumbing, and the prompt-versioning discipline that keeps a multilingual consumer agent from drifting once it is in production.
Who we would hire
People who have shipped an AI feature end-to-end and know which parts of the literature still apply to it. Taste over titles. Python or TypeScript at a strong level — we do not care which.
You would be the second engineer in. You would talk to users in week one. You would see your work on a customer's phone before you finish onboarding.
What you'll do
- Build and operate the shared agent stack used by all four ventures
- Own model selection, evals, and the prompt and data pipelines
- Talk to users from week one; ship work that reaches a customer's phone fast
- Pair with the founding team on product direction
What we're looking for
- Strong Python or TypeScript fundamentals — pick one
- Has shipped an AI feature in production (LLM, retrieval, voice, or vision — your call)
- Comfortable with evals, prompt versioning, and tool-use orchestration
- Singapore-based or willing to relocate
Apply for Founding AI Engineer
Send us the basics. A short note about why this role is interesting carries more weight than a long resume.