Confirmed on the employer's own hiring board on Nov 19, 2026. First seen by Alion on Nov 18, 2026.
Join our team as an AI/ML Software Engineer, where you will take AI use cases from an initial baseline to a customer-ready feature. You will work closely with the CTO, back-end and product teams, and a GRC domain expert. Your responsibilities will include building AI features integrated into the product, creating an AI gateway, developing document and retrieval pipelines, establishing evaluation and feedback loops, ensuring reliable production services, and laying the data foundations needed for these workflows. You should have at least 3 years of hands-on professional software engineering experience, strong Python skills, and experience deploying LLM or ML features to production. Fluency in French and English is required.
Missions
- Build AI features integrated into the product, starting with assisted drafting and expanding to evidence pre-review and campaign assistance as priorities allow.
- Create an AI gateway with approved model backends, tenant-specific access controls, budgets, fallbacks, and a tested switch-off mechanism.
- Develop document ingestion and retrieval pipelines that structure information while respecting tenant boundaries and existing permissions.
Profil recherché
At least 3 years of hands-on professional software engineering experience, building and shipping software to real production environments.
Strong Python software engineering: maintainable services, APIs, automated tests, debugging and code review.
You have personally deployed an LLM or ML feature to production for real users and supported it after release. You can explain your contribution, how you measured quality, and how you handled production failures.
Practical experience with LLM integration, structured outputs, embeddings and retrieval, including when a simpler deterministic solution is preferable.
SQL and data pipeline experience: ingestion, transformations, data quality and reliable operation.
Understanding of production trade-offs: latency, cost, retries, monitoring, access control and tenant isolation.
Ability to design a useful evaluation set, investigate failure cases and distinguish a measured improvement from noise.
Comfort working through an open problem with product and domain experts, delivering in small increments inside an existing platform.
Thoughtful use of AI coding tools such as Claude Code or Cursor, with ownership of the code and its verification.
We value evidence of work you have shipped over a particular degree or academic institution. Cybersecurity or GRC experience is helpful; you can learn the domain with our experts.
Languages: fluent french (much of the product, customer material and domain work is in French) and english

