Full Stack Software Engineer

About the role

ACENSI is seeking a Full Stack Software Engineer to join a client in the energy sector in Brussels. The role focuses on building AI‑centric applications that leverage large language models (LLMs), retrieval‑augmented generation (RAG), and agentic workflows. You will work hands‑on to deliver production‑grade solutions that are reliable, secure, and maintainable.

Key responsibilities

• Design and implement AI‑powered applications using LLMs, RAG, and agentic workflow patterns.
• Develop backend services, APIs, and integrations required for AI‑centric solutions.
• Build and optimise retrieval pipelines, including chunking, embeddings, vector and hybrid search, metadata filtering, and reranking.
• Work with LLM APIs and AI orchestration frameworks to create robust workflows.
• Implement evaluation, testing, monitoring, and observability for AI applications.
• Define safe patterns for tool use, human‑in‑the‑loop approval, and agentic behaviour.
• Collaborate with engineers, product owners, and business stakeholders.
• Select models and frameworks based on quality, cost, latency, maintainability, and security.
• Troubleshoot issues such as hallucinations, poor retrieval quality, high latency, and cost inefficiency.

Required profile

• Fluent in English and French or Dutch (spoken, read, written).
• Minimum 5 years of software engineering experience, preferably backend or full‑stack.
• 1–2 years of experience integrating LLMs or generative AI services.
• Practical knowledge of RAG, embeddings, vector search, and retrieval quality improvement.
• Strong programming skills in Python.
• Security‑conscious mindset when handling internal or sensitive data.
• Ability to communicate technical trade‑offs clearly and decide when AI is appropriate versus deterministic solutions.

Required skills

• Python
• Azure cloud environment
• LangGraph
• LangChain
• Semantic Kernel
• MicrosoftAgent Framework
• Microsoft Foundry
• Microsoft 365 Agents SDK
• CI/CD pipelines
• Testing and logging frameworks
• Vector search and embeddings
• Retrieval‑augmented generation (RAG) pipelines
• LLM APIs and AI orchestration frameworks
• MCP, A2A, tool calling, multi‑agent workflows