We’re seeking outstanding ML engineers and researchers to collaborate across multiple active Bradbury Group projects and papers.
Bradbury Group is a non-commercial, volunteer-only research lab pushing the frontier of open-source AI. Our work spans diffusion, model compression, geospatial, cancer research, and more; with one focus: producing conference-grade research that moves the field forward.
We are highly selective, and hold only a handful of open positions at any time. Your application will be kept for future sourcing.
What you’ll gain
- Coauthorship across multiple live papers and projects
- Access to funded compute for large-scale experiments
- Collaboration with world-class peers from top universities and labs
- A fully remote, async-friendly environment (volunteer/unpaid)
What we need
- Deep PyTorch and model design expertise; design and train custom models with clean, well-structured code.
- Familiarity with other common python libraries including
einops,OpenCV,pandas,matplotlib,hydra,wandb. - Strong ML fundamentals: transformers, ViT, diffusion, DiT/MMDiT, geometric ML, matrix math.
- Proven experience building real ML systems (beyond toy repos)
- Compute efficiency: profile bottlenecks, optimize compute, scale training and use budget smartly.
- Linux fluency: shell scripting, tmux/screen, perf tools.
- Disciplined workflows: clean code, async communication, clear documentation, small PRs, and careful use of LLM coding tools.
- Novel ideas and experimental, scientific thinking and testing.
Nice to have
- Leadership of multi-person teams or end-to-end projects.
- Experience with large compute, DDP/FSDP, Triton/CUDA.
- Domain expertise (geospatial, medical imaging, molecular biology).
- First-author publications or strong open-source track record.
- Strong drive and time-to-contribute.
Not required
- Specific timezone, visa, or work authorization
- PhD not required; some of our strongest contributors are exceptional undergraduates or industry engineers.