Position Overview
Autodesk Research brings together researchers, engineers, and industry domain experts in collaboration with top academic and industry partners. We are seeking a PhD intern to conduct cutting-edge research on test-time adaptation techniques for multi-modal foundation models in production environments. This unique role will collaborate closely with our ML engineering and research teams to develop novel approaches that enable robust model performance when deployed at scale with real-world data distribution shifts. This is an excellent opportunity for a doctoral candidate in Computer Science to work on impactful research with immediate real-world applications while contributing to publications and open-source projects.
Responsibilities
- Conduct literature surveys on test-time adaptions with foundation models, multi-modal learning, and related areas.
- Design and conduct experiments to evaluate adaptation performance across multiple modalities (vision, language, CAD, etc.), distribution shifts, and domain gaps, and identify opportunities for improvement.
- Collaborate with cross-functional teams to understand domain-specific challenges and integrate methods into real production workflows.
- Document research findings and present results to the broader research team and stakeholders
- Author research papers for submission to top-tier ML/AI conferences
Minimum Qualifications
- Currently enrolled in a PhD program in Computer Science, Mathematics, Electrical Engineering, or a closely related field, with demonstrated research experience with deep learning models, large-scale datasets, and transfer learning.
- Minimum 2 years of PhD-level research experience
- Strong publication records in top-tier peer-reviewed conferences and journals
- Proficiency in Python, deep learning frameworks (PyTorch, Lightning, TensorFlow) and libraries (Transformers, Scikit-Learn, Matplotlib)
- Basic understanding of distributed computation on GPU clusters and hands-on experience with schedulers (SLURM, Kubernetes, Ray, etc.)
- Basic knowledge with AWS services (S3, SageMaker, etc.)
- Solid understanding of deep learning fundamentals and LLM fundamentals
- Strong analytical and problem-solving skills with the ability to work independently on ambiguous research problems
- Excellent written and verbal communication skills for presenting complex technical concepts
Preferred Qualifications
- Prior research experience specifically with test-time adaptation, domain adaptation, or continual learning
- Prior research experience with CAD data
- Experience with transformer models, multi-modal models (e.g., CLIP)
- Track record of first-author publications at top-tier ML conferences (NeurIPS, ICML, ICLR, CVPR, ICCV, SIGGRAPH, etc.)
- Experience with fine-tuning machine learning models on cross-domain data
- Experience in monitoring machine learning models in production environments
- Familiarity with concepts such as efficient fine-tuning or parameter-efficient adaptation techniques
- Strong software engineering practices and experience with version control (Git)
About the Canada Intern Program
The 2026 Canada Internship program runs for 16 weeks (May 4th – August 21st). All internships are paid. As an intern, you will contribute to meaningful projects, be mentored by industry leaders, and participate in tech talks and other activities designed to support your personal and professional development. Our internships align with Autodesk’s Flexible Workplace approach, which is designed to meet the needs of our business while providing flexibility in support of office, remote and hybrid work preferences.
