Open Positions
All roles are part-time volunteer; 3 roles currently open.
Group-Wide ML Expectations
We group all machine-learning roles under a single category because most of our work spans multiple projects and demands cross-project responsibility. The expectations below apply broadly to all ML roles, with reasonable flexibility for positions that are primarily engineering-focused or rely on specialised domain knowledge (e.g., protein-modeling expertise).
Required
- •Team-based ML experience. Has worked as part of a research lab or engineering team (academic or industry), with code review, iteration cycles, and shared responsibilities.
- •Real project experience beyond toy work. Has completed ML projects involving non-trivial datasets, non-trivial training, or multi-week experimentation - not just notebooks or coursework.
- •Experience implementing non-standard architectures. Has reproduced or adapted ideas from papers (custom attention blocks, diffusion variants, special-purpose modules, etc.) rather than only using standard libraries.
- •Independent contributor with no need for hand-holding. Able to pick up a task, define the missing pieces, and drive it to completion.
- •Strong PyTorch experience.
- •Strong experimental discipline. Understands baselines, controls/ablation, reproducibility, basic tracking tools, and proper evaluation.
Preferred
- •Meaningful contribution to a published or submitted paper (ICLR/CVPR/NeurIPS/CVPR or equivalent).
- •First-author experience on a peer-reviewed or arXiv-preprint research project.
- •Deep familiarity with diffusion, transformers, flow matching.
- •Designed or proposed your own architectures or training schemes, not purely reproductions.
- •Current PhD or completed MSc student in a relevant field, or equivalent research experience. (We strongly acknowledge that skilled undergraduates often outperform more senior candidates, this is not a rigid selection criteria).
Nice to Have
- •Completed a PhD, especially in ML, applied math, CV, or scientific ML.
- •Experience with large-scale compute environments: SLURM, multi-GPU training (DDP, FSDP)
- •First Author top ML conference paper (ICLR/CVPR/NeurIPS etc.)
- •Prior work on niche domains (e.g., protein modelling, medical imaging, generative editing, scientific ML).
- •Triton / CUDA kernels.
- •Leadership experience / interest in first author / project management roles.
◈
Select a role to view details
Open Positions
All roles are part-time volunteer; 3 roles currently open.
Group-Wide ML Expectations
We group all machine-learning roles under a single category because most of our work spans multiple projects and demands cross-project responsibility. The expectations below apply broadly to all ML roles, with reasonable flexibility for positions that are primarily engineering-focused or rely on specialised domain knowledge (e.g., protein-modeling expertise).
Required
- •Team-based ML experience. Has worked as part of a research lab or engineering team (academic or industry), with code review, iteration cycles, and shared responsibilities.
- •Real project experience beyond toy work. Has completed ML projects involving non-trivial datasets, non-trivial training, or multi-week experimentation - not just notebooks or coursework.
- •Experience implementing non-standard architectures. Has reproduced or adapted ideas from papers (custom attention blocks, diffusion variants, special-purpose modules, etc.) rather than only using standard libraries.
- •Independent contributor with no need for hand-holding. Able to pick up a task, define the missing pieces, and drive it to completion.
- •Strong PyTorch experience.
- •Strong experimental discipline. Understands baselines, controls/ablation, reproducibility, basic tracking tools, and proper evaluation.
Preferred
- •Meaningful contribution to a published or submitted paper (ICLR/CVPR/NeurIPS/CVPR or equivalent).
- •First-author experience on a peer-reviewed or arXiv-preprint research project.
- •Deep familiarity with diffusion, transformers, flow matching.
- •Designed or proposed your own architectures or training schemes, not purely reproductions.
- •Current PhD or completed MSc student in a relevant field, or equivalent research experience. (We strongly acknowledge that skilled undergraduates often outperform more senior candidates, this is not a rigid selection criteria).
Nice to Have
- •Completed a PhD, especially in ML, applied math, CV, or scientific ML.
- •Experience with large-scale compute environments: SLURM, multi-GPU training (DDP, FSDP)
- •First Author top ML conference paper (ICLR/CVPR/NeurIPS etc.)
- •Prior work on niche domains (e.g., protein modelling, medical imaging, generative editing, scientific ML).
- •Triton / CUDA kernels.
- •Leadership experience / interest in first author / project management roles.
◈
Select a role to view details
Open Positions
All roles are part-time volunteer; 3 roles currently open.
Group-Wide ML Expectations
We group all machine-learning roles under a single category because most of our work spans multiple projects and demands cross-project responsibility. The expectations below apply broadly to all ML roles, with reasonable flexibility for positions that are primarily engineering-focused or rely on specialised domain knowledge (e.g., protein-modeling expertise).
Required
- •Team-based ML experience. Has worked as part of a research lab or engineering team (academic or industry), with code review, iteration cycles, and shared responsibilities.
- •Real project experience beyond toy work. Has completed ML projects involving non-trivial datasets, non-trivial training, or multi-week experimentation - not just notebooks or coursework.
- •Experience implementing non-standard architectures. Has reproduced or adapted ideas from papers (custom attention blocks, diffusion variants, special-purpose modules, etc.) rather than only using standard libraries.
- •Independent contributor with no need for hand-holding. Able to pick up a task, define the missing pieces, and drive it to completion.
- •Strong PyTorch experience.
- •Strong experimental discipline. Understands baselines, controls/ablation, reproducibility, basic tracking tools, and proper evaluation.
Preferred
- •Meaningful contribution to a published or submitted paper (ICLR/CVPR/NeurIPS/CVPR or equivalent).
- •First-author experience on a peer-reviewed or arXiv-preprint research project.
- •Deep familiarity with diffusion, transformers, flow matching.
- •Designed or proposed your own architectures or training schemes, not purely reproductions.
- •Current PhD or completed MSc student in a relevant field, or equivalent research experience. (We strongly acknowledge that skilled undergraduates often outperform more senior candidates, this is not a rigid selection criteria).
Nice to Have
- •Completed a PhD, especially in ML, applied math, CV, or scientific ML.
- •Experience with large-scale compute environments: SLURM, multi-GPU training (DDP, FSDP)
- •First Author top ML conference paper (ICLR/CVPR/NeurIPS etc.)
- •Prior work on niche domains (e.g., protein modelling, medical imaging, generative editing, scientific ML).
- •Triton / CUDA kernels.
- •Leadership experience / interest in first author / project management roles.
◈
Select a role to view details
Socials
© 2025 Bradbury GroupLegal
Socials
© 2025 Bradbury GroupLegal