
I'm a final year Ph.D. candidate in Computer Science and Engineering at the University at Buffalo, where I design and build robust systems for real-time cyber-physical applications. My Ph.D. research focuses on the "systems" dimension of Visual SLAM -- specifically, how system design can drive the performance and reliability of Visual SLAM in real-world environments.
My work sits at the intersection of systems research and applied engineering. I have extensive experience with a wide range of technical stacks and robotics/vision libraries, including Rust, C++, Python, ROS, OpenCV, GTSAM, g2o, and Foxglove. A core theme of my research is introducing Rust into the robotics space. Most recently, I led the development of Bauhaus, a Rust-based framework for modular, memory-safe Visual SLAM systems.
Outside academia, I’ve contributed to production-scale systems at Google and edX -- from building predictive models to reduce Chromium page latency, to improving video infrastructure and engagement features in large-scale learning platforms.
Publications
-
Sofiya Semenova, Steven Ko, Yu David Liu, Lukasz Ziarek, and Karthik Dantu. A Comprehensive Study of Systems Challenges in Visual Simultaneous Localization and Mapping Systems. ACM Transactions on Embedded Computing Systems (TECS), August 2024.
-
Ali J. Ben Ali, Marziye Kouroshli, Sofiya Semenova, Zakieh Sadat Hashemifar, Steven Y. Ko, and Karthik Dantu. Edge-SLAM: Edge-Assisted Visual Simultaneous Localization and Mapping. ACM Transactions on Embedded Computing Systems (TECS), August 2022.
-
Sofiya Semenova, Steven Y. Ko, Yu David Liu, Lukasz Ziarek, and Karthik Dantu. A Quantitative Analysis of System Bottlenecks in Visual SLAM. HotMobile 2022. Best Poster Award
-
Ali J. Ben Ali, Sofiya Semenova, and Karthik Dantu. Platform Variability in Edge-Cloud Vision Systems. HotMobile 2019.
-
Ryan Marcus, Olga Papaemmanouil, Sofiya Semenova, and Solomon Garber. NashDB: An Economic Approach to Fragmentation, Replication and Provisioning for Elastic Databases. SIGMOD 2018.
-
Ryan Marcus, Sofiya Semenova, and Olga Papaemmanouil. A Learning-based Service for Cost and Performance Management of Cloud Databases (Demonstration). ICDE 2017.
Projects
-
For older projects, see the projects page of this website.