Role Overview
You will lead the architecture, training, and deployment of large language models powering a widely used AI companionship platform. With tens of millions of daily tokens processed, your work will directly influence the scalability, safety, and responsiveness of the core product. This is a hands-on leadership role where technical depth meets real-world impact in a fully remote setup.
Key Responsibilities
- Lead by doing—design systems and mentor engineers while remaining deeply involved in the Python/PyTorch codebase
- Own the end-to-end chat experience, optimizing context handling, memory retrieval, and response latency
- Define strategies for fine-tuning and preference alignment, including SFT, RLHF, and DPO
- Make critical decisions on when to use prompting, fine-tuning, or retrieval-augmented generation
- Oversee the lifecycle of training data—from sourcing and labeling to cleaning—for improved model behavior across cultures
- Build intelligent content filters that detect harmful inputs without compromising user experience in an explicit content environment
- Develop context-sensitive moderation systems that go beyond simple flagging to understand nuance
What We’re Looking For
You have a proven track record of deploying machine learning systems at scale, particularly with LLMs serving millions of users. You bring 8+ years of engineering experience with deep fluency in Python, PyTorch, and modern AI tooling like vLLM and HuggingFace. You understand how model parameters influence behavior—temperature, sampling, and prompt design shape the personality of the system.
Equally important is your mindset: you prioritize shipping over perfection, obsess over metrics and regressions, and take full ownership of the code you write. You’re comfortable navigating the complexities of uncensored content and know how to balance openness with safety through technical rigor.
Work Environment & Benefits
- Fully remote role, open to candidates in France
- Flexible contract options with a preference for B2B arrangements
- 20 days of paid time off annually
- Annual in-person gathering to connect with the team
- Monthly health insurance stipend and access to mental health support for you and family members
- Co-working and learning budgets to support your productivity and growth
- Equipment support including a company laptop, monitor allowance, and premium tool access (ChatGPT, Cursor, Claude, Hugging Face)
Culture & Team
We operate with high autonomy and minimal bureaucracy. You’ll have direct access to the CTO and work alongside a distributed team of top-tier engineers who value speed, ownership, and creative problem-solving. Success here means long-term impact, technical excellence, and building systems that scale responsibly.
