Founding ML engineer
Le poste
Preferred Experience
🔍 Who we're looking for
5+ years of experience in Machine Learning Engineering or software development and data architecture.
Strong coding experience using Python or Typescript and Rust
Expertise in deep learning frameworks: PyTorch, JAX, or TensorFlow.
Experience in Inference and RL framework: vLLM, SGLang, Slime
MLOps & model deployment know-how: versioning, monitoring, experiment tracking.
Proven track record delivering projects in an agile, collaborative environment.
Genuine desire to mentor and grow more junior profiles over time.
Demonstrated ability to thrive in a demanding, fast-growing environment.
Excellent analytical skills, attention to detail, and strong prioritization.
Fluent English, mandatory (working language with our clients and the labs).
The extras that make the difference
Experience with GCP (or equivalent), Modal or GPU cluster and serverless technologies
Experience with infrastructure as code (Terraform, CloudFormation).
Strong interest in data anonymization, multi-modality, and RL environments.
🎁 Why join us
Entrepreneurial adventure, play a key role and live the early days of an ambitious company, right at the front line of the AI revolution.
Backed by Y Combinator, join just as we accelerate.
Founding impact, shape the product, the tech stack, and the engineering culture from day one.
Offices in central Paris + 1–2 days WFH.
-
Package:
- €70–90k/year depending on profile
- BSPCE 0.2–0.5% depending on profile
- Alan health insurance (mutuelle)
- 50% Navigo covered
Recruitment Process
A 30-min intro call with Grégoire (COO & co-founder)
A 30-min call with Thomas (CTO & co-founder)
-
A technical case study:
- Async preparation
- Debrief interview with Thomas
A final fit interview with the founding team
Offer
Additional Information
- Contract Type: Full-Time
- Start Date: 21 September 2026
- Location: Paris
- Occasional remote authorized
- Salary: between 70000€ and 90000€ / year
Ooak data
Ooak Data construit Alexandria, une bibliothèque de jeux de données représentant des workflows réels d'entreprises. La société acquiert des environnements de données complets, les anonymise et les structure afin qu'ils puissent servir à l'entraînement et à l'évaluation des agents d'intelligence artificielle.
L'objectif est de confronter les agents à la complexité des outils, documents et processus réellement utilisés par les entreprises, au-delà des benchmarks académiques. Ces environnements sont destinés aux laboratoires d'IA qui développent des modèles capables d'accomplir des tâches professionnelles concrètes.
Équipe
5 personnes
Ville
Paris