Lightberry
The social brain for robots.
Software Engineer, Voice AI
About the role
Humanoid robots doing backflips are super cool, but to be useful they need to be able to listen, communicate, and operate in the world. Your role in our organization would be to help make that a reality by developing the cloud-based side of our robots that makes them intelligently understand the people interacting with them.
What you'll do
- Use robots that cost more than most cars to evaluate and test new features
- Iterate on software deployed to customer's robots in the real world
- Improve our composite voice pipelines using the latest state of the art models
- Add and improve features like knowledge storage/retrieval, smarter turn taking, automatic personality tuning
- Create clever approaches to reducing latency and make interactions feel more natural
What we’re looking for
- Experience with cloud development (AWS, GCP, Kubernetes)
- Experience with latency sensitive applications (voice pipelines, video games, streaming)
- Experience with voice AI
- Excited to take high ownership of product features
- Addiction to observability metrics
- Excellent taste in developer tooling
Technology
- Our cloud code is primarily Python 3.14 and it uses a custom UDP-based streaming protocol to communicate with the robots
- We deploy at edge locations across the world in GCP and AWS
About Lightberry
Lightberry develops software to make any robot emotionally intelligent and socially aware.
Our robots are learning to behave naturally around humans: they listen, speak, see, and navigate the world like we do.
We work with multiple robot makers, including the #1 humanoid robot manufacturer in the world, Unitree. Lightberry is backed by top funds including Y Combinator, Liquid2, a16z Scout, and operators at NVIDIA, Meta, Google, and more.
Our software is currently state of the art in human-robot interactions.
