Postdoctoral Researcher in Knowledge and Data Engineering
Join a cutting-edge research project on multimodal foundation models, diffusion models, and large-scale AI systems.
Best for:Best for researchers with a background in deep learning, NLP, or ML systems who are eager to work on large-scale generative AI models.
Overview
The Knowledge and Data Engineering Laboratory at Toyota Technological Institute is seeking a postdoctoral researcher to contribute to a major NEDO-funded project focused on multimodal foundation models. The successful candidate will work on foundational technologies for next-generation language generation, including diffusion and flow-based models. The role offers access to high-performance computing resources, including the AI Bridging Cloud (ABCI), and encourages publication in top-tier international conferences. Applicants may focus on either model analysis and learning algorithms or large-scale training and inference systems. This position provides a collaborative environment with opportunities to engage with industry partners and international research networks.
What you'll do
- Conduct foundational research on pre-training, distillation, and optimization of large-scale diffusion and flow-based language models.
- Analyze internal mechanisms, interpretability, and generation dynamics of generative models to improve performance.
- Implement and optimize large-scale training and inference pipelines using multi-GPU and multi-node distributed systems.
- Develop and publish high-quality research in top-tier international conferences such as ACL, EMNLP, NeurIPS, or ICLR.
- Collaborate with project partners and contribute to open-source software development where applicable.
- Participate in regular research discussions and present findings within the laboratory and at academic conferences.
Requirements
Degree required: PhD in Computer Science, Engineering, or a related field (or expected completion by start date).
Language: English or Japanese; proficiency sufficient for academic communication and documentation.
Instruction: English and Japanese
Publications: Strong preference for candidates with a track record of publications in top-tier international conferences.
- Proven experience in developing deep learning models using PyTorch.
- Solid foundational knowledge of Large Language Models.
- Strong proficiency in written and spoken English or Japanese.
- Ability to conduct independent research and contribute to collaborative projects.
- Must hold or be expected to hold a PhD degree by the start date.
- Must be able to commit full-time to the NEDO project.
Funding & terms
How to apply
Deadline: 2027-03-31 · open until filled
Interested candidates should submit a CV, a list of publications, a summary of past research, a research plan, and one letter of recommendation via email. Informal online interviews are highly encouraged prior to formal application to discuss research fit and environment.
- CV
- List of publications
- Summary of past research
- Research plan
- 1 reference letter
Contact
Create a free profile to unlock your personal match score, the full requirements, funding details and how to apply — in English or Persian.