LLM Engineer

About us

We’re working on Sky: natural-language computing for your Mac.
Read more about the team, our values, and our vision at www.software.inc/jobs

The Role

We’re looking for an LLM engineer who will play a core role in shaping our product. You will be directly responsible for designing our data pipelines, fine-tuning architecture, and evaluation systems. You’ll make make important tradeoffs between ambition and feasibility. And you’ll run a tight loop between fine-tuning → evaluation → deployment.

Every day you will...

...build unprecedented software. You'll use your experience and your enthusiasm to turn big ideas into concrete plans. Sometimes this means you'll have to start running before you can see the finish line. But your ability to learn matters more than your existing knowledge.

...take radical ownership for your projects. You will be directly responsible for the success of your systems and features. You'll do what it takes to get a project done. You're always looking to incorporate feedback and improve quality. We trust you with big responsibilities.

...influence architecture decisions. You'll use your knowledge of the latest model architectures to design systems for data collection, training, and inference. You'll optimize model performance to strike a strategic balance between latency and helpfulness. You intuitively find ways to prioritize both user privacy and robust data collection.

...think big and small. You understand that your infrastructure choices have important impact on the end-user experience. With an eye on performance, you'll create large, robust models for cloud inference or precise, on-device models where every megabyte matters.

Qualifications

• Bachelors' degree in Computer Science, or equivalent experience building software

• Experience writing code to integrate models into production environments

• Experience implementing and deploying systems for data collection, cloud model training and inference.

• Experience translating current, relevant AI research into production systems.

This is a full-time hybrid role. We work together in-person in San Francisco, at least 3 days per week.