AI Research Scientist, Multimodal
Autodesk
PythonPyTorch
AI Research Scientist, Multimodal VLM and MLLM
Autodesk
full-time
Posted on:
Origin: • 🇬🇧 United Kingdom
Visit company websiteJob Level
Mid-LevelSenior
Tech Stack
PythonPyTorch
About the role
- Do fundamental and applied research to help customers imagine, design, and make a better world
- Develop new ML models and AI techniques (including multimodal VLMs and MLLMs)
- Lead on research projects within a global team
- Review relevant AI/ML literature to identify emerging methods, technologies, and best practices
- Explore new data sources and discover techniques for best leveraging data
- Collaborate with scientists, researchers, engineers, and designers across projects in computer vision, graphics, robotics, HCI, sustainability, simulation, manufacturing, architectural design and construction
- Report to a Research Manager in the AI Lab
Requirements
- A Master's or PhD in a field related to AI/ML (Computer Science, Mathematics, Statistics, Physics, Linguistics, Mechanical Engineering, or related disciplines)
- Strong background applying Deep Learning techniques (custom architectures, optimizing model performance, novel loss functions, metrics, benchmarks, deploying production-ready solutions)
- Familiarity in statistical methods for Machine Learning (e.g. Bayesian methods, HMMs, graphical models, dimension reduction, clustering, classification, regression techniques)
- Familiarity with PyTorch, PyTorch Lightning, or similar frameworks
- Strong coding abilities in Python
- Strong software development skills and familiarity with GPUs, MLOps, Git, high-performance large-scale ML systems and platforms
- Experience with Transformers, LLMs, NLP, Multi-modal Deep Learning, and VLMs/MLLMs
- Publication track record in machine learning conferences and/or journals
- Preferred: significant post-graduate research experience, or five or more years of work experience
- Preferred: experience with foundation and open-source models, meta/transfer learning, fine-tuning, distillation, multi-task/continual learning, RL, RLxF, and/or image/audio/video/3D generation/understanding
- Preferred: experience or familiarity with 3D Graphics/Animation, Manufacturing, Engineering, Architecture, and/or Construction domain
