ML Engineer (AI)

TL;DR

ML Engineer (AI): Building and optimizing an on-device inference engine for Apple Silicon, contributing to model optimization and proprietary runtime. Focus on designing evaluation metrics, building data processing pipelines, and experimenting with model architectures and training methods.

Location: Remote (global). Onsite options available in San Francisco or Europe.

Company

Mirai builds the fastest on-device inference engine for Apple Silicon, outperforming MLX and llama.cpp on supported models with a small, senior, and deeply technical team.

What you will do

• Design evaluation metrics and benchmarks to assess model performance.
• Build data processing and labeling pipelines for model training.
• Experiment with model architectures and various training methods.
• Contribute to the core training and model optimization framework.

Requirements

• Strong background in Deep Learning with deep knowledge of modern text foundation models.
• Solid grasp of statistics and probability theory.
• Understanding of the ML performance optimization landscape.
• Experience building complex training pipelines from scratch.
• Very strong software engineering skills.

Nice to have

• Experience with JAX.
• Experience with GPU programming.
• Knowledge of Rust.
• Thorough understanding of concepts like muP, Muon, FSDP, FlashAttention, and SigLIP.

Culture & Benefits

• Work in a small (14 people), senior, and deeply technical team.
• Opportunity to ship fast and own problems end-to-end.
• Advised by a former Apple Distinguished Engineer.
• Backed by leading AI-focused funds and individuals.
• Work on making local inference practical, fast, and reliable for real products.