The role

You'll help lead the effort to architect, train and deploy machine learning models that enable Sidekick to excel in a variety of tasks related to scene understanding, code generation and more.

This is a builder’s role. You'll have the opportunity to take on ambiguous problems and deep AI research and translate them into a platform that removes the distraction and tedium from building games.

Sidekick is the game development counterpart. It uses the project, game engine documentation, and best practices to provide contextual suggestions and execute changes directly within the engine. It helps debug projects, iterate on scenes, and execute standardizable tasks, so small teams can iterate faster and spend more time on the creative parts of game development. Sidekick currently supports Unity and will support other popular game engines in the near future.

You may be a fit if…

• You’re excited about pushing beyond RAG-based pipelines and into a much more advanced system utilizing agents and custom models

• You live in Python and/or JavaScript and Typescript

• You have experience conducting multi-GPU, distributed model training, and fine-tuning on remote compute clusters

• You have strong knowledge of at least one machine learning framework (e.g. PyTorch, Tensorflow, JAX)

• You excel at setting up, running, and optimizing ML experiments

• You have a Masters or PhD Degree in Computer Science, a related technical field, or equivalent practical experience

• You stay up-to-date on the latest research and think of creative ways to integrate it in a product excites you

Nice to haves include…

• You enjoy writing about the work you do. You may already have publications in a top machine learning, computer vision, or computer graphics conference (e.g NeurIPS, ICML, ICLR, CVPR, SIGGRAPH)

• You’re familiar with game engines like Unity or Unreal

About Bezi

Bezi is the team behind Sidekick. Our mission is to empower the people behind the games. To do this, we build products that work with designers and developers - enabling them to be more creative and efficient.

We’re a group of technical, first-principles thinkers. Every member of the team is trusted, and expected, to deeply understand the people we build for and think creatively about ways to improve their lives and work. Our organization is flat; everyone is trusted to pursue and validate ideas and every idea is supported and debated equally.

We believe the best work happens in teams that know each other. Bay Area team members gather in our SoMa office twice a week.

How to apply

Send your resume and a short note/link of a project you’re proud of to careers@bezi.com. If there’s a potential fit, our team will respond as soon as possible.