Senior AI Engineer

The Senior AI Engineer owns the enterprise LLM platform substrate that powers every generative-AI consumer across the organisation. This role designs, builds, and operates the LLM Gateway, the evaluation framework, and the AI best-practices playbook that engineers, business teams, and product squads depend on to ship trustworthy AI features at scale.

Key Responsibilities

• Design, build, and operate the enterprise LLM Gateway — provider abstraction, authentication, rate limiting, per-use-case cost tracking, prompt logging for audit, and model routing across major providers (Azure OpenAI, AWS Bedrock, Anthropic, Vertex AI).
• Operationalise an evaluation framework (Langfuse or equivalent) — tracing, eval scores, human feedback loops — as a platform service consumed by every production GenAI consumer.
• Define and enforce evaluation rubrics (accuracy, groundedness, hallucination rate, latency, cost per inference) and embed regression gates in CI so no GenAI consumer ships without a working eval harness.
• Ship and operate the GenAI cost dashboard with per-use-case attribution and quarterly forecasts to leadership; drive cost-optimisation initiatives (caching, prompt compression, model routing).
• Partner with Data Governance on AI-model governance evidence (audit-log schema, PII redaction proofs, model routing controls) to support legal and compliance approvals for GenAI at scale (Thailand PDPA, regional AI frameworks).
• Partner with platform engineering on Vector Search and embedding model selection, retrieval relevance tuning, chunking strategies, and reranker layers.
• Author and publish the AI Best Practices Playbook — the standard engineering teams across the company use when shipping LLM features; mentor and review.
• Own GenAI platform service-level objectives — availability, latency, cost ceilings — and lead incident response for production GenAI consumers.

• Bachelor's or Master's degree in Computer Science, AI/ML, Data Science, or a related discipline.
• 6+ years of software engineering with at least 2+ years shipping production LLM platform systems (gateways, evaluations, cost metering, multi-model routing).
• Strong Python production-service development (FastAPI, async, observability, tests).
• Hands-on production experience with at least one major LLM provider (Azure OpenAI, AWS Bedrock, Anthropic, Vertex AI), including cost and latency optimisation.
• Eval-driven LLM development discipline — golden sets, LLM-as-judge, regression gates in CI, multi-step conversation replay.
• Solid grounding in prompt-injection defence, data-leakage prevention, and PII handling for LLM systems.
• Cloud production experience (Azure preferred; AWS/GCP transferable) and Git-based CI/CD; comfortable owning service contracts and SLOs.
• Excellent written and verbal communication; able to author technical design documents and influence engineering peers and business stakeholders.

Preferred Qualifications

• Experience with LLM gateways (Portkey, LiteLLM, Kong AI) or having built one in-house.
• Fine-tuning experience (LoRA / QLoRA) and open-weight model deployment (vLLM, TGI).
• Thai-language NLP exposure (PyThaiNLP, WangchanBERTa, SEA-LION, Typhoon) and retail / commerce data context (POS, catalog, CRM, loyalty).
• Vendor or industry certifications such as Databricks Generative AI Engineer, Azure AI Engineer Associate, or comparable.