Generative AI LLM Engineer – Automotive Software

About the role

Micla Engineering & Design is building a team for a major international vehicle manufacturer to embed generative AI across its software development, testing, and quality engineering processes. This position focuses on designing, deploying, and optimizing large‑language‑model (LLM) based systems that work directly with engineering artefacts such as requirements, test cases, code, and release documentation.

Key responsibilities

• Design, deploy and fine‑tune generative AI and LLM solutions for software‑engineering workflows.
• Architect agentic systems that orchestrate tool integration, task decomposition and multi‑step execution.
• Build retrieval‑augmented generation pipelines using embeddings, vector stores and grounding strategies over engineering artefacts.
• Create prompts and structured output schemas (e.g., JSON schema, function calling) that produce machine‑verifiable results.
• Develop evaluation layers, including correctness benchmarks, regression suites, safety guardrails and drift detection.
• Operate LLMOps pipelines: versioning, monitoring, cost and latency management.
• Integrate AI‑assisted steps into existing CI/CD pipelines with proper quality gating.

Required profile

• Engineering degree in Computer Science, Software Engineering, Data Science, Electrical Engineering or a related field.
• Strong Python programming skills.
• Hands‑on experience building production applications on top of LLM APIs or open‑weight models.
• Practical knowledge of retrieval‑augmented generation techniques (embeddings, vector stores, chunking, grounding).
• Experience with prompt engineering, structured output design and tool‑using agents.
• Proficiency with Git and integrating work into CI/CD pipelines.
• Professional written and spoken English.

Required skills

• Python
• LLM APIs / open‑weight models
• Embeddings and vector stores
• Prompt engineering and JSON schema design
• Git
• CI/CD platforms (Jenkins, GitLab CI, TeamCity)
• Agent frameworks and orchestration tooling
• LLMOps observability and evaluation harnesses
• Model fine‑tuning, adapters, distillation