AI Scientist, Agentic Engineering
Vor Ort · Paris (Frankreich) · Senior · Vollzeit
Mehr Stellen wie diese?
Erstelle dein kostenloses Profil bei PyHires: Du siehst, wie gut du zu jeder Stelle passt, und bekommst jede Woche eine E-Mail mit den Stellen, die zu dir passen.
Überblick
- Arbeitsmodell
- Vor Ort
- Standort
- Paris (Frankreich)
- Anstellung
- Vollzeit
- Erfahrung
- Senior
- Sprachen
- Englisch
- Offen für Bewerber/innen aus
- Nur EU-Länder
- Unternehmen
- Mistral.ai
- Veröffentlicht
- 19. Nov. 2026
Über die Stelle
Mistral builds frontier models and full-stack AI solutions. In its AI4Engineering Science team, AI Scientists expand what Mistral's agentic tools can do across the engineering lifecycle in CAE (CFD, FEA) and EDA/semiconductors. Your core work is building the pre- and post-training data that lets Mistral's LLMs reason about and execute real engineering tasks, and shaping the agent architectures that make them reliable in multi-step workflows.
What you'll do
- Build pre- and post-training data for real engineering tasks
- Define verifiers and evaluation criteria that capture what correct, high-quality work means for each task
- Design and improve agent architectures and harnesses: planning, tool calls, error recovery and long-horizon chaining
- Build evaluation benchmarks and diagnostic tooling, and trace failures to gaps in data, reward design or scaffolding
- Work with CAE and EDA domain experts to choose the highest-value workflows to target
- Contribute to Mistral's wider pre- and post-training research
What you bring
- Deep, hands-on machine learning expertise, especially LLM development
- Experience running, debugging and validating real engineering workflows in CAD, CAE, semiconductor simulation or EDA
- Clean, readable Python and comfort in Linux/HPC environments
- Fluent English and the ability to explain ML and engineering concepts to any audience
- Self-direction and a collaborative, low-ego attitude
Nice to have
- Building or fine-tuning agentic systems
- Reward modelling, RLHF/RLAIF/RLVR or preference-based training
- Industrial or academic experience with CAE or EDA tools, and publications at engineering and ML venues
What they offer
- A benefits package that varies by country (healthcare, parental leave, retirement plans, relocation support)
Anforderungen
Basiert auf:Python
Muss
- LLMUnverzichtbar
- Linux
Über Mistral.ai
Mistral provides full-stack AI solutions, from frontier models to developer tools, applications and compute. It partners with enterprises in high-stakes industries such as finance, manufacturing, defence, healthcare and the public sector to build customised AI systems they can run on their own terms, with teams across Europe, North America, Asia and the Middle East.
Stack
- Python
- LLM
- Linux
Unternehmen
Mistral.ai
Frankreich