Dr. Aisha Rahman
AI Research Scientist | NLP & Multimodal Learning
Summary
AI Research Scientist with a Ph.D. in natural language processing and 6 years of research experience. Author of 15 peer-reviewed papers (2,400+ citations) spanning LLMs, multimodal learning, and model interpretability. Passionate about reproducible, open research that transfers to real systems.
Experience
Research Scientist at Neuralis Institute for AI
Jul 2021 to PresentSan Francisco, CA
- Led a team of 5 researching efficient fine-tuning of large language models, publishing 4 papers at ACL and NeurIPS.
- Developed an interpretability toolkit for transformer models used by 30+ academic labs.
- Released 3 open-source model checkpoints downloaded over 2M times.
Postdoctoral Researcher at Veritas Labs
Sep 2019 to Jun 2021Toronto, Canada
- Researched multimodal representation learning, contributing to a model that set SOTA on 3 benchmarks.
- Mentored 4 graduate students, co-authoring 6 publications with them.
Education
Ph.D. in Electrical Engineering & Computer Science at Massachusetts Institute of Technology
Sep 2015 to Jun 2019Cambridge, MA
M.Sc. in Computer Science at University of Toronto
Sep 2013 to Jun 2015Toronto, Canada
Technical Skills
Core: LLMs, Transformers, Multimodal Learning, Interpretability, Efficient Fine-tuning, Evaluation, Tools: PyTorch, JAX, Hugging Face, Weights & Biases, Ray, Git
Selected Publications
Efficient Adapters for Multimodal Transformers | NeurIPS 2024 (poster, 11% acceptance)
2024Interpreting Attention in Large Language Models | ACL 2023 (long paper)
2023Cross-Modal Retrieval with Shared Embeddings | ICML 2022
2022Talks, Awards & Service
Invited Talk: Scaling Evaluation for Modern LLMs | AI Frontiers Conference
2024Best Paper Runner-Up | Workshop on Efficient ML, ICML
2023