Software Engineer
Nous Research
Software Engineer
Nous Research
Open-source AI research enabling decentralized training
If you're a great engineer with a wide-ranging skill set, we want you to apply and join the team. You'll help us build Hermes Agent, Nous Portal, and other infrastructure that powers Nous Research. A lot of our team is made up of generalist engineers, and people like this tend to fit really well here.
Product surfaces people touch every day across Hermes Agent and Nous Portal, on web, desktop, and mobile.
Backend services and APIs that hold up under real usage.
The cloud platform and infrastructure behind our products and model serving.
Developer and research tooling that makes the rest of the team faster.
Whatever the highest-leverage problem is that quarter, which is not always predictable at a company moving this fast.
A few years building and shipping production software that real people depend on.
Depth in at least one area (frontend, backend, infrastructure, systems, or ML tooling) and the range to work outside it.
Fluency in the languages we use day to day, mainly TypeScript and Python, and comfort reaching for a systems language like Rust when it's the right call.
Strong ownership: you see something broken or missing and you fix it without waiting to be asked.
Heavy use of AI-assisted development in your own workflow.
Clear communication and low ego, because the team is small and everyone reads each other's code.
Experience with AI products, LLM applications, agents, or tool-calling workflows.
Open-source contributions, or side projects you're proud enough to show us.
Familiarity with modern AI infrastructure: PyTorch, OpenAI-compatible APIs, model routing, or inference.
Open-source AI research enabling decentralized training
Nous Research creates open-source AI models and simulators using a decentralized, community-driven approach that coordinates contributors worldwide to share idle GPU power. Its Hermes language models, along with tools like Psyche Network, DisTrO, DataForge, and Atropos, enable training and development through a token-based system on the Solana blockchain. This setup emphasizes transparency, auditable training workflows, and greater model steerability compared to many commercial systems, supported by open-source infrastructure. The goal is to democratize access to advanced AI by building world-class, open models and a scalable, transparent decentralized training platform anyone can contribute to or use.
Explore more opportunities at Nous Research