R&D Engineer
Ultralytics · Remote · Remote · Madrid ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- Fully remote
- R&D Engineer
- PyTorch expert
- Computer Vision
- Equity package
Join the team behind the world's leading open-source YOLO models to shape the future of vision AI. You will operate at the intersection of research and production, designing novel architectures and optimizing performance for millions of users.
This position requires deep expertise in PyTorch and computer vision, offering the chance to build original models from scratch rather than just fine-tuning existing ones. You will collaborate with a global R&D team to define future releases across the Ultralytics ecosystem.
What you'll do
- Design and build novel neural network architectures for next-generation YOLO models
- Develop core components such as backbones, necks, heads, and attention modules
- Implement cutting-edge research papers into working model architectures from scratch
- Run rigorous ablation studies and large-scale experiments on standard benchmarks
- Improve model accuracy and latency through distillation, pruning, and quantization
- Partner with engineering teams to convert research ideas into production-ready code
What you'll need
- 5+ years of hands-on experience in computer vision and deep learning
- Proven success building models from the ground up, not just fine-tuning
- Expert-level Python skills and deep proficiency in PyTorch
- Strong command of convolutions, attention, normalization, and loss design
- Ability to read research papers and implement them faithfully from scratch
- Experience with efficiency research like quantization, pruning, or NAS
- Portfolio of research contributions via papers, preprints, or open-source code
Nice to have
- Published research at venues such as CVPR, ICCV, ECCV, NeurIPS, or ICLR
- Experience with distributed training, mixed precision, or CUDA profiling
- Active open-source presence with architecture-focused repositories
- Familiarity with the Ultralytics Python package documentation or community
- History of taking original research from prototype to widely used models
Stack
- Python
- PyTorch
- YOLO
- CUDA
How you'll work
- Fully remote position
- Flexible working hours focused on results
- Global collaboration with teams in London, Madrid, and Shenzhen
What you get
- Competitive salary reflecting experience and contribution
- Equity packages for team members
- 24 vacation days plus birthday off and local holidays
- Brand-new Apple MacBook Air/Pro, Studio Display, and AirPods Pro
- Dedicated budget for personal and professional development