Job Specifications
Formel AI
At [Formel AI](https://formel.ai/), we combine the creativity of generative AI with the reliability of mathematical formal methods. By leveraging the verification capabilities of formal methods, such as the Lean 4, we pioneer a new generation of Large Language Models (LLMs) that are reliable, transparent, and cost-efficient. We're honored to have received the [2nd spot of the Innovation Prize from École normale supérieure](https://www.linkedin.com/posts/ecolenormalesuperieure_jeudi-soiront-%C3%A9t%C3%A9-annonc%C3%A9s-les-premiers-ugcPost-7401284253482635264-odbQ) in Nov 2025.
Formel AI is founded by [Sylvain Combettes](https://www.linkedin.com/in/sylvain-combettes/) (CEO) and [Antoine Mazarguil](https://www.linkedin.com/in/antoine-mazarguil/) (CTO), who met during their PhDs at ENS Paris-Saclay. We closed an [oversubscribed angel round](https://www.linkedin.com/posts/sylvain-combettes_with-antoine-mazarguil-were-thrilled-activity-7429435501692006400-i3u0) in Feb 2026, backed by 30+ top-tier business angels including the founders of Datadog, Dataiku, Nabla, and Artefact. Our scientific advisory board includes Full Professors from Ecole polytechnique, ENS Paris-Saclay, and Ecole des Ponts.
We're now onboarding our first design partners and building the team to make it happen. Come join us!
Adresse
Formel AI
75000 Paris
France
Détail de l'offre (poste, mission, profil)
Corps de texte
About The Role
As our AI Engineer, you will be at the forefront of AI innovation, working on groundbreaking technology that combines large language models with formal verification methods. You'll join a founding team of deep-tech experts tackling one of the most critical challenges in AI: building LLMs that are transparent, reliable, and cost-effective.
This is a hands-on research and engineering role where you'll design and develop the next generation of formally verified AI systems. You'll work on novel architectures, orchestration protocols, and training methodologies that eliminate hallucinations while dramatically reducing computational costs.
You'll be joining at a critical inflection point—moving from research breakthroughs to real-world deployment with design partners in high-stakes domains like healthcare and enterprise software, where AI reliability is non-negotiable.
Key responsibilities
Design and develop AI systems: Lead the architecture and implementation of LLMs integrated with formal verification methods
Build orchestration protocols: Design and implement orchestration frameworks, including tool calling, structured generation, and verification workflows
Drive research innovation: Actively shape the company's research roadmap, exploring new architectures, training methodologies, and verification techniques in collaboration with our scientific advisory board
Deploy with design partners: Work closely with 2-3 early customers in high-stakes domains to implement proof-of-concept solutions that demonstrate measurable improvements in AI reliability and cost-efficiency
Optimize training and evaluation: Develop advanced dataset generation pipelines, fine-tuning workflows (DPO, RLHF), and rigorous evaluation protocols to continuously improve model performance
Build reusable infrastructure: Create documentation, code templates, and engineering best practices that enable the team to scale our deployment approach as we grow
Bridge theory and practice: Translate cutting-edge research in AI into production-grade systems
What we are looking for
Essential qualities
Research-driven curiosity: You're passionate about exploring novel AI applications and architectures, pushing the boundaries of what's possible
Hands-on technical excellence: You can design, implement, and deploy production-grade AI systems autonomously, from model training to orchestration
Deep technical expertise: You have deep knowledge of LLM architectures, training methodologies, and inference optimization
Bridge builder: You can translate complex research concepts into practical implementations.
Background
3+ years of hands-on experience building and deploying AI/ML systems, with deep expertise in Python, LLM APIs, and cloud infrastructure
Strong foundation in machine learning fundamentals, model architectures, and modern AI engineering practices
Deep knowledge of the latest developments in AI research and production systems
Required Skills
Python mastery: You write clean, efficient Python code and understand object-oriented programming principles deeply
Fullstack fundamentals: You understand both frontend and backend development principles, can work across the stack when needed, and are proficient in Python for backend systems
LLM expertise: You have hands-on experience with large language models, including API integration, prompt engineering, and orchestration frameworks
LLM inference packages: You have practical experience with at least one major LLM inference framework (vLLM, llama.cpp, TensorRT-LLM, or similar) and understand in
About the Company
Créé en 2017 dans le cadre de la Stratégie Nationale pour l’Intelligence Artificielle, l’Institut DataIA est le pôle d’excellence en IA de l’Université Paris-Saclay. Il fédère 14 établissements d’enseignement supérieur et de recherche, dont CentraleSupélec et l’ENS Paris-Saclay, ainsi que des organismes nationaux et partenaires académiques. L’Institut œuvre à structurer l’écosystème IA autour de la recherche, de la formation et de l’innovation, avec des actions phares comme le projet SaclAI-School, lauréat de l’appel Compét...
Know more