Senior AI Engineer — RAG Applications

pSTIB-MIVB (Brussels public transport operator) is building several employee
- and customer-facing AI services, serving our customers and employees across the organization. The AI services are build and ran on our STIB AI Platform. /ph3Your role /h3pWe are recruiting a Senior AI Engineer to join the Digital Innovation / AI CoE team for the Build and hypercare phases of some our most complex AI services. You will work alongside STIB's internal AI Platform Architect, the other Senior AI Engineer (already on board), and the project teams (project manager, business analyst, solution architect), turning our AI projects into running, evaluated and maintainable services. /ppThe AI Platform runs on a hybrid operating model: managed Azure services (Azure AI Foundry with the OpenAI model family for LLM and embeddings, Azure AI Search, Azure API Management, Azure AI Content Safety) combined with self-hosted open-source components on AKS (agent orchestration in Python with LangChain, ingestion pipeline), with MLflow on Databricks for evaluation and tracing. The Digital Innovation / AI CoE team owns Infrastructure-as-Code for all application-scoped Azure resources, using Terraform and Azure DevOps pipelines. /ph3What you will deliver /h3pActivities will vary from project to project. Typically you will deliver: /pullithe RAG pipeline end-to-end, from retrieval strategy to answer generation ; /lilithe ingestion pipeline: connectors to source platforms, chunking, embedding, incremental refresh ; /lilisecurity trimming and ACL propagation ; /liliagent orchestration, prompt templates and guardrails ; /lilithe evaluation and tracing setup: offline evaluation sets, retrieval and answer-quality metrics, end-to-end tracing of requests and agent steps ; /lilithe Terraform, Helm and Azure DevOps artefacts needed to deploy and operate the service ; /liliObservability and cost control: instrumentation for latency, token consumption and run cost, plus quality regression detection after model or prompt changes ; /liliHypercare and early production: incident analysis, tuning, runbooks and documentation, and handover of the service to the engineer who will own the Run phase. /li /ulh3Roles /h3pSenior Artificial Intelligence (AI) Engineer /ph3Skills /h3ulliLangChain /liliPython /liliRAG (Retrieval-Augmented Generation) /liliTerraform /li /ulh3Languages /h3pDutch or French (native or bilingual proficiency), English (full professional proficiency) /ph3Questions /h3ulliHave you personally built or led the implementation of a Retrieval-Augmented Generation (RAG) system that went to production? (if answer to Q1 is "Yes") Please describe the RAG system briefly: business domain, approximate document corpus size, retrieval stack used (embeddings, vector store, reranking if any), and the retrieval quality metrics you measured. Maximum 300 words. /liliHave you personally authored Kubernetes manifests or Helm charts and Terraform code for Azure, as part of deploying an AI or ML workload to production? (if answer to Q2 is "Yes") Please give one concrete example where you were personally responsible for both the Kubernetes (manifests or Helm charts) and the Terraform code. Maximum 200 words. /liliHave you designed and implemented an evaluation framework (metrics, test sets, automation) for a RAG or LLM application you worked on? (if answer to Q3 is "Yes") Please describe the evaluation methodology: which metrics you used, how you curated or generated test sets, and how evaluation was integrated into your deployment or iteration cycle. Maximum 200 words. /liliHave you personally implemented document-level access control in a RAG system, so that users only see content they are authorised to access in the source platform? (if answer to Q4 is "Yes") Briefly describe how permissions were propagated from the source system to the retrieval layer, and how you tested it. Maximum 200 words. /liliThis role requires on-site presence in Brussels 2 to 3 days per week. Are you willing and able to commit to this schedule? /li /ul