About the role
Hume AI is a startup focused on developing emotionally intelligent AI, with a mission to enhance human well-being through speech and audio understanding models. The engineering team is dedicated to creating systems that support the deployment of trained models into production, ensuring they operate efficiently and reliably. The company emphasizes scientific research and ethical guidelines in AI development, fostering an innovative environment in New York City.
The role involves owning and optimizing the entire inference pipeline, from model checkpoints to live deployment. Responsibilities include managing graph export, engine compilation, runtime integration, and building scalable serving platforms. The engineer will work closely with research and backend teams to improve system performance, correctness, and operational safety, with a high degree of autonomy in system design and implementation.
Candidates should have extensive experience in systems software, Linux, and at least one systems programming language like C++, Rust, or Go. Knowledge of neural network execution, distributed systems, and profiling tools is essential. The role is onsite, full-time, and involves diagnosing production issues, optimizing performance, and developing infrastructure for serving multiple models across products.