Divyanshu Saxena

I am a graduate student at the Department of Computer Sciences, University of Texas at Austin. I am fortunate to be advised by Prof. Aditya Akella at the UT Networked Systems (UTNS) lab. I also collaborate closely with researchers at the LDOS project where we are building next-generation Operating Systems that can leverage learning to improve performance and adaptability.

I am broadly interested in Systems and Networking. My Ph.D. research has been on developing techniques to achieve Performance Robustness for computer systems by combining ideas from Machine Learning, Optimization, Programming Languages and Formal Methods.

Previously, I was an undergraduate student at the Indian Institute of Technology, Delhi majoring in Computer Science and Engineering where I met and made friends with some absolutely amazing and competent people. I hail from Lucknow, a city in Northern India.

CV:   One-pager   Long

news

Sep 24, 2025 Our paper on joint learning for hardware caching and prefetching accepted to appear at the ML for Systems Workshop at NeurIPS ‘25!
Sep 16, 2025 Our framework for using LLMs to generate instance-optimal policies, PolicySmith, accepted to appear at HotNets’25!
Aug 23, 2025 Our paper on integrating verification in the learning of congestion controllers is accepted to appear at EuroSys’26!
Jul 1, 2025 Invited to serve on the Program Committee for ACM Symposium on Cloud Computing 2025! Check out the Call for Papers here.
Jun 2, 2025 Excited to join the Systems Innovation group at Microsoft as a Research Intern this summer!
Mar 31, 2025 Our paper on Guardrails for learned OS policies accepted to appear at HotOS’25!
Oct 4, 2024 Copper Wire, our work on expressive and efficient service mesh policies, is accepted at ASPLOS’25!