Hi! I am Bakshree Mishra, a final-year PhD candidate in the Department of Computer Science at the University of Illinois Urbana-Champaign, advised by Prof. Sarita Adve. I work on computer architecture and ML systems.

I am on the job market for full-time research and engineering roles starting in 2027. Here is my CV.

My research focuses on designing scalable and energy-efficient systems for rapidly evolving application domains, especially artificial intelligence (AI) and extended reality (XR). Traditional approaches to accelerating critical computation with hardware specialization rely on stable adoption of algorithms given the long life-cycle of chip design; however, state-of-the-art algorithms change rapidly in these domains. My work seeks to answer how to build and repurpose scalable systems that can enable users to run workloads across various domains despite changes to algorithms. I investigate this question in three directions:

  • Composable accelerators: trading off the degree of hardware specialization by composing smaller accelerators in a shared-memory SoC (Mozart, PACT 2024).
  • Scaling LLM inference on 3D-stacked GPUs: identifying and addressing network-on-chip (NoC) bottlenecks in large 3D-stacked GPU-memory architectures, in collaboration with AMD Research (under review).
  • Distributed edge inference: reusing accelerators across edge devices as a distributed system to improve the latency and energy efficiency of local inference (preliminary results at MLArchSys 2025; full paper under review).

My ongoing work brings these directions together to orchestrate multiple AI models across edge and cloud.

Before my PhD, I worked at Intel in Bangalore, first as a Graduate Technical Intern and then as a Machine Learning Engineer and an IP Logic Design Engineer. I designed FPGA accelerators for computer vision and ML algorithms, co-designed accelerators for recommender systems and graph neural networks, and designed front-end RTL for SoC IPs. During my PhD, I have also worked with AMD Research in Bellevue (2024 and 2025) on data movement and LLM inference in 3D-stacked GPU-memory architectures.

I received my Master’s in Computer Science from the National Institute of Technology, Rourkela, where my research on real-time pedestrian detection systems was advised by Prof. Bansidhar Majhi, and my undergraduate degree from the Odisha University of Technology and Research (formerly College of Engineering and Technology), Bhubaneswar.


Updates

Jan 2026: I passed my Preliminary Examination! Many thanks to my advisor, my committee, and my research group for their guidance and support!
Dec 2025: I completed the CRA UR2PhD Graduate Student Mentor Training Course to learn how to be a better mentor, and I learnt a lot. I highly recommend the course!
Aug 2025: I am continuing with AMD Research part-time as a Research Associate through November 2025!
July 2025: One of the patent applications I filed with my colleagues at Intel, Register fault detector, was granted US Patent No. 12,373,295. Congratulations to my colleagues and thank you for teaching me so much in this project!
June 2025: Gave an oral presentation on Minstrel: Application-Aware SLM Inference Optimization on Edge Devices at the Workshop on ML for Computer Architecture and Systems (MLArchSys), co-located with ISCA 2025!
May 2025: I am returning to AMD RAD, Bellevue, for a second summer internship!
June 2024: Our work, "Taming the Acceleration Tax: Enabling New Opportunities for Fine-Grained, Disaggregated Accelerator-Level Parallelism", got accepted into PACT 2024 (published as Mozart: Taming Taxes and Composing Accelerators with Shared-Memory)! That's over 2 years hard work, systems design and research on actual hardware. Many congratulations to my co-authors and especially grateful for my advisor's faith in our work!
May 2024: I will be interning with the RAD team at Bellevue this summer! I am excited to work with John Alsop, Michael Boyer and Mike Ignatowski on tackling interesting research problems!
Feb 2024: Our paper on a 12nm Linux-SMP-capable RISC-V SoC with 14 accelerator types, a multi-institution collaboration, appeared at ISSCC 2024! [Paper]
Nov 2022: I passed my oral Qualifying Examinations Fall '22! Many thanks to my advisor and my colleagues in my lab and research group for their guidance and support!
Sep 2022: I was listed among Teachers Ranked as Excellent for my stint as a TA for CS225 (Introduction to Data Structures and Algorithms with C++). Thank you!
Jun 2022: Our patent application on "Graph Reordering and Tiling Techniques", filed along with my amazing colleagues+co-inventors at Intel, Bangalore, got published! [Pub. No. US202/015632A1]
May 2022: Attended the ADA Symposium 2022 at University of Michigan, Ann Arbor.
Apr 2022: Attended the Grad Cohort - Women workshop organized by Computing Research Association, where I got to interact with peers and leaders in Computer Science Research.
Aug 2021: August 4th was my last day at Intel, Bangalore. Thank you for a great learning experience! Also excited to share that I will be joining the University of Illinois, Urbana Champaign later this month to pursue a PhD in Computer Science!
Jun 2021: Filed two patents applications with co-inventors at Intel! Pub.Nos.: US2021/0319022A1 and US2021/031832A1
Apr 2021: Presented poster and paper at Intel HSPE Techbuzz '21 (Intel-internal conference), and won best paper in track! Thank you!
Jan 2021: Was awarded Prolific Innovator Award (Intel Division Recognition Award) at Intel HSPE TechBuzz Q1 '21. Thank you!
Dec 2020: Named as one of the Top 50 Intel India employee volunteers of 2019! Thank you!
Dec 2019: Received Intel Departmental Recognition Award for category Spirit of HSPE! Thank you!
Nov 2019: Paper on our work on hardware acceleration of ML and CV algorithms got accepted at EAI Endorsed Transactions on Cloud Systems! [Paper]
Sep 2019: Presented our paper on "Edge Acceleration of Computer Vision and Deep Learning Algorithms on the Edge using OpenCL" at IEEE WinTechCon 2019, Bangalore! Talk slides can be found here.
Jun 2019: Received Intel Departmental Recognition Award for my contributions to setting up demos at Customer Experience Center at Intel Bangalore. Thank you!
May 2019: Presented live demonstration of our work on custom accelerators for machine learning algorithms at Intel DTTC 2019 at Portland, OR!
Apr 2019: Two papers and a live demo were selected for Intel Design & Test Technology Conference (DTTC) 2019. I will be traveling to Portland, Oregon, to present our demo!
Jun 2017: Won 2nd runners' up in Intel India WIN Hackathon on Inclusion Day with our solution to help the visually challenged: NAINA - Alarm Identification and Navigation Aid. Thank you!
Jun 2017: Joined Intel as Machine Learning Engineer.
May 2017: End of internship at Intel and culmination of my Master's study at NIT Rourkela. My thesis was titled "Real-Time Pedestrian Detection System Using OpenCL-Based FPGA Acceleration".
May 2016: Moved to Bangalore and joined Intel as Graduate Technical Intern to conduct research on hardware exploration of compute intensive algorithms
Jul 2015: Joined NIT Rourkela for Master's studies, advised by Prof. Bansidhar Majhi