Research Engineer - Tokenizer & Protocols
Paris • Full-time
Posted 9d ago
Remote Work Policy
On-site
Employment Type
Full-time
Categories
LLM Engineer
About the job
Mistral is seeking a Research Engineer to own and evolve the core libraries for tokenization and model interfaces that underpin their full-stack AI solutions. This role is critical as it impacts the entire AI stack, from pretraining to deployment, ensuring consistency in how training, evaluation, and production systems communicate. You will translate research needs into robust, shared tooling, working across the company's frontier models and developer tools.
Responsibilities
- Design, build, and maintain core libraries for tokenization and model interfaces.
- Evolve formats and tokenizers to adapt to changing models and use cases.
- Translate research requirements into robust, shared tooling.
Requirements
- Strong software engineering background with excellent Python skills.
- Excellent understanding of tokenizers, including BPE, special tokens, chat templates, and tokenizer training.
- Hands-on experience with tokenization libraries like tiktoken or Hugging Face tokenizers.
- Experience building production-grade libraries used by many individuals.
- A strong focus on correctness and attention to detail.
- Good knowledge of LLM training and serving processes.
- Clear communication, low ego, and a collaborative mindset.
Benefits
- Healthcare coverage
- Parental leave
- Retirement plans
- Relocation support
- Wellness programs
- Meal and transportation allowances
