Akash Pandey

Akash Pandey

PhD Candidate, Northwestern University

Akash Pandey

I am a Ph.D. candidate in the Department of Mechanical Engineering at Northwestern University. I am being co-advised by Dr. Sinan Keten and Dr. Wei Chen.

Currently, I am working on accelerating the mutation-based primary sequence optimization in proteins by taking into account the contribution of each monomers in the sequence. In the past, I have worked on explainable AI models for proteins to understand the contribution of each monomers in the sequence, predicting protein’s dynamic behavior and mechanical properties.

While working on DL models for proteins, I have developed a special interest in sequence-based DL models. I really enjoy architecting and training DL models from scratch. Due to this interest, I have worked on projects/challenges involving biosignals and audio signals too. In June 2023, I and my other teammate secured the third position in one of the ICASSP’23 challenges and presented that work in the conference. One of my other papers on emotion share prediction using large language model embeddings has been published in ACM MM’23 as one of the first authors.

During June-August 2024, I worked as a Data Scientist PhD Intern at Capital One. During the internship, I worked on Explainable AI methods for transformer-based models to estimate feature importance as well as the importance of different time points in a sequential data. Prior to joining Northwestern, I worked as a Lead and Advanced Engineer at Infosys and Rolls-Royce respectively. In both companies, I was responsible for the stress analysis and fatigue life assessment of Titanium and Nickel alloy-based rotating discs in Rolls-Royce Trent-XWB engines. I have a Masters by Research in Applied Mechanics from Indian Institute of Technology, Madras during which I worked on characterizing the fatigue properties of smart piezoelectric composite material using experimental as well as Finite Element Analysis techniques. 

You can reach me at akash.pandey@northwestern.edu
Link to my CV

Recent Highlights

  • [March,2025] - Our paper titled “COLOR: A compositional linear operation-based representation of protein sequences for identification of monomer contributions to properties” is accepted at MLGenX workshop in ICLR 2025.
  • [Feb,2025] - Submited our paper on surface-EMG based silent-speech recognition using LLMs to ACL.
  • [Jan,2025] - Our novel eXplainable AI (XAI) method, named COLOR, developed to estimate contribution of each monomer in the primary sequence is now online on arXiv.
  • [June-August,2024] - Completed 10 weeks Data Science PhD internship at Capital One, San Jose.
  • [May, 2024] - Successfully completed my PhD thesis proposal.
  • [May, 2024] - Paper titled “Sequence-based data-constrained deep learning framework to predict spider dragline mechanical properties” published in Nature Communication Materials.
  • [Sept, 2023] - Recipient of Predictive Science and Engineering Design (PSED) fellowship.

Services

  • Guided Xiaoyuan Zhang, a Master student in EECS department of Northwestern University, to develop a Llama (2 and 3) based deep-learning model to predict text from silent EMG signals on a closed vocabulary.
  • Guided Yueyuan Sui, a Master student in EECS department of Northwestern University, for ACM MM 2023 challenge for human emotion prediction task.
  • Reviewer at Nature Computational Materials and IEEE/ACM Transactions on Computational Biology and Bioinformatics.