Bangkok, Thailand (Hybrid)
Position: Platform Engineer
Location: Bangkok, Thailand
Work model: Hybrid
Employment type: full time.
RESPONSIBILITIES:
Develop and maintain AI agents so employees across departments can use them in real work, including onboarding new users, adding tools/integrations, fixing issues and gathering feedback to iterate
Continue developing the AI Gateway, adding new data sources, writing and adjusting OPA policies, expanding the retriever and maintaining audit logs, following default-deny and least-privilege principles
Improve the RAG pipeline by ingesting new documents, improving embedding/retrieval quality, labelling data sensitivity and measuring answer accuracy
Maintain LLM serving, managing vLLM/Ollama on GPU, evaluating and testing new models, managing context length and memory budgets, and integrating external APIs responsibly (keys, cost, rate limits)
Build automated AI workflows for business teams (e.g. report summarisation, support assistance, data analysis), taking projects from a proof of concept through to a maintainable production service
Own work end-to-end - design, build, test, deploy and support after go-live - following the team's tech standard, and track usage metrics for everything shipped to report progress consistently
Write documentation and train employees to use AI tools correctly and safely
REQUIREMENTS:
2+ years of programming experience with strong Python skills (the company's systems are primarily Python-based) and comfortable working on Linux/command line
Fluent in English both verbal and written
Hands-on experience building real things with LLMs (beyond using ChatGPT), such as calling APIs, building RAG, building agents/tool calling, or fine-tuning, with deep understanding of at least one of these
Solid understanding of LLM fundamentals: tokens, context windows, embeddings, prompt engineering, hallucination and mitigation strategies
Understanding of AI security, including prompt injection, data leakage, and why an authorisation layer between AI and company data is necessary
Experience using AI coding tools (Claude Code, Codex, Copilot, etc) as part of daily workflow
Fast learner, comfortable working with unfamiliar systems, and able to read others' code/documentation and build on it
OFFER:
Hybrid model
Attractive remuneration, based on skill set and experience
Hands-on work with real GPU servers, real models and real users
Clear growth path towards becoming lead engineer of the company's entire AI platform
And many more!