Required skills
- Python
- Google Cloud Platform
- Vertex AI
- LLMOps
- Prompt Engineering
- Google Agent Development Kit
- LangGraph
- Cloud Run
- Google Kubernetes Engine
- BigQuery
- Pub/Sub
- LLM Security
Languages: English · French
Job description
Project overview
Are you an experienced Lead AI Engineer with production GenAI and multi-agent experience on Google Cloud?
emagine has a longterm opportunity for you to drive the design and production rollout of a multi-agent GenAI system with a global team.
Skills & Experience
Tasks & Deliverables
Why join
Other details
Apply with your latest CV or reach out to Brendan to find out more.
- 5+ years in software, data or ML engineering, including 1-2+ years with LLM or agent solutions live in production (not only POCs)
- Expert Python (typed, tested, maintainable) and expert GCP: Vertex AI (Gemini, Agent Engine, RAG Engine, Gen AI Evaluation, Vector Search), Cloud Run/GKE, BigQuery, Pub/Sub, IAM, VPC-SC
- Advanced agentic architectures (orchestrator/sub-agents, planner-executor, human-in-the-loop), function calling and memory management, with Google ADK, LangGraph or equivalent
- Expert context and prompt engineering, plus LLMOps: evaluation datasets, LLM-as-a-judge, tracing, token FinOps
- LLM security (prompt injection, OWASP Top 10 LLM, guardrails) and production run experience (incidents, SLOs, monitoring)
- Nice to have: GitHub Actions, Terraform, Docker, MCP, A2A, AI Act and responsible AI, Google Cloud Professional certification
- Professional English (B2 minimum, C1 desired) alongside French for daily work
- Run discovery with operational business teams: challenge requests, prioritise use cases by value, feasibility and risk, and define success and adoption metrics
- Propose the target architecture and build the foundation: agents, tools, memory, evaluation, observability and security
- Prototype quickly with users, then move to production and keep the system robust in run
- Optimise quality, cost and latency, and keep answers reliable, traceable and explainable
- Coach the team through code reviews and pair programming, and create reusable patterns for other agentic AI products
- A long-term mission, from MVP build through industrialisation and scale-up to run
- Roughly 70% build and innovation work on cutting-edge agentic AI
- A multidisciplinary agile team of 6-10 (product owner, data scientists, data and software engineers)
- High visibility: your architecture becomes the reference for future projects
- Contract: 2-3 years
- Location: Hybrid, with 2 days per week on site in Roissy (Paris-CDG), the rest remote
- Working style: Collaborative, enabling, and we value autonomous, self-starting candidates
Project details
- Start: flexible
- Duration: > 12 months
- Remote: hybrid
- Location: Tremblay-en-France, France, FR
- Languages: English, French
Indexed by 9am's meta-search. You are a job provider? Claim or manage this listing
Found on portal.emagine.org