資深 AI 工程師 | Senior AI Engineer (LLM & Agent)

【About the Role】
Build and operate JKOPay's AI Foundation Platform and customer-facing AI features. You'll own the end-to-end delivery of LLM applications, agent workflows, and RAG systems — designing for both internal product teams and end users, in a PCI DSS-compliant financial services environment.

【Why This Role】
You'll work on AI infrastructure that ~70 engineers depend on daily, and product features that reach millions of JKOPay users — with real ownership, modern tooling, and the technical freedom to make AI work in a regulated environment.

【What You'll Do】

• Ship LLM applications: prompt engineering, RAG pipelines, function calling, and multi-agent orchestration powering real product features
• Operate the AI platform: extend our LiteLLM Gateway + Langfuse + pgvector stack on GKE; build evals, observability, and CI/CD for prompts and models
• Own AI safety & governance: design data boundaries within PCI DSS scope, implement guardrails (PII detection, prompt injection defense), and support compliance audits
• Choose the right model for the job: balance self-hosted vs. API, cost vs. latency, quality vs. throughput
• Partner across teams: work with Product, Backend, Data, and InfoSec to embed AI into the product and internal workflows

【What We're Looking For】

• 3+ years of software engineering, with at least 1.5 years building LLM / AI applications
• Hands-on with RAG (vector DBs, chunking, retrieval tuning), agent frameworks (LangGraph, CrewAI, or equivalent), and at least one major LLM provider API(Anthropic, OpenAI, Google)
• Proficient in at least one programming language and able to write production-quality code. We work primarily in Python for LLM development; strong engineers from Kotlin / Java / Go / .NET backgrounds who can ramp on Python are welcome
• Comfortable with Kubernetes, Docker, and CI/CD; familiarity with GCP or AWS
• Experience with LLM observability tools (Langfuse, LangSmith, Phoenix, or similar)

【Nice to Have】

• AI experience in finance, e-commerce, or other regulated industries
• Working knowledge of PCI DSS, GDPR, or data protection law in the context of AI
• Hands-on experience with prompt injection defense, jailbreak mitigation, or PII handling
• Familiarity with MCP / A2A / ACP or other agent interoperability protocols
• Fine-tuning (LoRA / QLoRA) or self-hosted model deployment (vLLM, TGI, Ollama)
• Open-source contributions or technical writing in AI / ML