Applied Scientist — Multimodal AI & Research Engineer

Mistral; New York (NY); On-site; USD 150,000 - 210,000; Full time

Benefits offered by this job

Healthcare coverage

Relocation support

Meal and transportation allowances

Well-being programs

Job summary

Mistral AI is seeking Applied Scientists and Research Engineers to drive innovative research across modalities such as text, image, and speech. You will develop novel methods, apply models to diverse enterprise use cases, and collaborate with internal and client teams to deliver high-impact AI solutions.

You will work with cross-functional teams to tackle complex problems, build scalable tooling, and contribute to a large codebase while maintaining high standards of code quality and reliability.

Qualifications

• Excellent communication skills and ability to explain complex concepts to diverse audiences.
• Expertise with PyTorch or JAX and strong Python coding skills.
• Able to contribute to large codebases with minimal guidance.

Responsibilities

• Run pre-training, post-training, and deployment of state-of-the-art models on large GPU clusters.
• Generate and curate data for pre-training and evaluation; ensure performance exceeds expectations.
• Develop tools and frameworks for data generation, training, evaluation, and deployment.
• Collaborate across cross-functional teams on complex use cases using agents and RAG pipelines.
• Manage research projects and stakeholder communications.

Skills

PyTorch or JAX

Python

Communication

Education

PhD or Master in ML/CS

Tools

Codebase familiarity

Job description

Mistral AI is seeking Applied Scientists and Research Engineers to drive innovative research across modalities such as text, image, and speech. You will develop novel methods, apply models to diverse enterprise use cases, and collaborate with internal and client teams to deliver high-impact AI solutions.

You will work with cross-functional teams to tackle complex problems, build scalable tooling, and contribute to a large codebase while maintaining high standards of code quality and reliability.