Atos is hiring an AI Delivery Architect to help define and deliver AI architecture patterns, engineering standards, evaluation frameworks, and governance practices across Belgium delivery teams. The role combines strategic architecture ownership with hands-on technical delivery, supporting production AI systems from concept to scale.
Key Responsibilities
- Define and maintain AI delivery reference architectures, pattern libraries, and target-state architectures for AI engagements.
- Establish engineering standards for LLM applications, agent orchestration, prompt management, retrieval pipelines, model serving, observability, and lifecycle management.
- Work directly with delivery teams to translate strategy into shipped artefacts, including Python MCP servers, typed agent boundaries, and evaluation gates in CI.
- Conduct AI risk assessments covering EU AI Act classification, provider-versus-deployer determination, ISO 42001, DORA, NIS2, and GDPR alignment.
- Evaluate AI platforms, model providers, sovereign vendors, agent platforms, evaluation tooling, and open-source components based on fit, cost, performance, sovereignty, and risk.
- Partner with platform, MLOps, and LLMOps teams on CI/CD, infrastructure-as-code, observability, drift detection, incident response, and AI FinOps.
- Drive responsible AI practices, including privacy, auditability, bias mitigation, explainability, human-in-the-loop controls, and technical documentation for high-risk classifications.
Required Qualifications
- Bachelor’s or Master’s degree in Computer Science, Engineering, Data Science, or a related field, or equivalent senior engineering experience.
- 8+ years in software or enterprise architecture, distributed systems, or AI/ML platform engineering.
- 3+ years of hands-on experience designing and shipping production AI/ML or generative AI systems, including LLM systems taken from prototype to production.
- Strong hands-on engineering skills in Python, MCP servers, typed agent boundaries, evaluation harnesses, and CI integration.
- Production experience with AWS, Azure, or Google Cloud, and at least one sovereign-EU stack.
- Deep knowledge of LLM application architecture, including RAG, function calling, agentic orchestration, fine-tuning, prompt orchestration, and model serving.
- Familiarity with MLOps and LLMOps tooling, model lifecycle management, observability frameworks, and CI/CD for AI systems.
- Working knowledge of EU AI Act risk classification, ISO/IEC 42001, DORA, NIS2, GDPR, enterprise security, IAM, encryption, data protection, and data-residency requirements.
- Ability to communicate architecture decisions clearly to delivery teams and senior stakeholders.
Preferred Qualifications
- Experience in regulated industries such as financial services, healthcare, public sector, or critical infrastructure.
- Relevant certifications, architecture framework experience, internal developer platform or AI accelerator experience, and public writing, open-source contributions, or conference talks related to AI engineering systems.
Technical Skills
AI/ML: LLMs, RAG, embeddings, vector search, model evaluation, fine-tuning, agentic orchestration, and classical ML. Agent platforms include Anthropic Claude, Microsoft Copilot Studio and Agent 365, AWS Bedrock AgentCore, Google Vertex AI Agent Builder, and Atos Polaris. Architecture and engineering skills include distributed systems, microservices, event-driven architecture, API design, Python, FastAPI, async, Pydantic, pytest, TypeScript or Go, CI/CD, and infrastructure-as-code. Data and platform skills include Databricks, Snowflake, vector stores, knowledge graphs, AWS, Azure, GCP, Kubernetes, Docker, serverless, GPU infrastructure, and sovereign stacks. Observability, operations, governance, and security skills include Langfuse, Arize Phoenix, Ragas, DeepEval, Promptfoo, OpenAI Evals, monitoring, tracing, drift detection, AI FinOps, IAM, zero-trust, EU AI Act, ISO 42001, DORA, and responsible-AI guardrails.