Job Description
We are looking for someone to join our company and push the frontiers of enterprise AI further.
You will join our Bangkok team, building a new insurance platform for a large insurer. The engineers who do well here own their work end to end: understanding the problem, deciding what to build, writing it down clearly, and building it.
Most of the work right now is writing functional specifications. You take the client's business requirements, decide what the system actually has to do, and write it down clearly enough for their team to approve. That means reading the existing codebase, working out what is realistic to build, and defending those decisions in front of the client.
This project is not typical of what we do, and when it ends you will join our other engineers working on production AI systems for large organisations in regulated industries.
Being a startup, direction and execution clarity are often vague, iterative, and less structured. If you thrive in a place where autonomy given is high, and are able to translate broad visions to concrete steps, this is for you.
What You'll Do
- Turn the client's business requirements into clear, detailed, testable functional specifications
- Study the codebase closely enough to judge what is realistic, and raise concerns before the specification is written
- Take part in client workshops, follow up on open questions, and document what was agreed
- Write and review production TypeScript as development work comes to us
- Move onto our other engagements and our own products, building production AI systems for enterprises in regulated industries
- Work alongside our core and research teams on retrieval, LLM orchestration, and on-premise deployment
Must Have
- 2 to 4 years of professional experience. If you have less but the rest of this list fits you, we would still like to hear from you
- Production TypeScript, or strong fundamentals in another typed language. You will need it to read the system before you need it to write
- React, Node/Express, REST APIs, Git, Docker and Linux at a working level
- Production experience in more than one language or framework, and the ability to explain what you changed in your approach when you moved
- The habit of asking what a requirement means before building against it
- Professional English and native Thai. Client workshops are conducted in Thai, and our team works in English
Nice to Have
- Something you built because you wanted it to exist, rather than because it was assigned to you
- A codebase you took over from someone else and improved
- Freelance or client work where you were the only technical person involved
- Competitive programming, open source, or research carried out outside your job
- Python, or experience with LLM or AI agent systems
- Insurance, banking, or government systems, or work on the vendor side of a large client project
About Pints AI
Pints AI is an award-winning AI company based in Singapore. We build on-premise, privacy-preserving AI systems for some of Asia's most regulated industries, including banks, insurers and government agencies.
- Won the AI in Finance Global Challenge (2023) at SFF in Singapore
- Built 1.5 Pints, a compact open-source LLM that outperforms Apple's OpenELM and Microsoft's Phi-1.5
- Built Autothought, our AI framework engine for back office automation in regulated industries
- Built SentientRAG, a low-hallucination RAG framework for compliance-heavy data

