Prateek Ganguli

Research Interests

My research focuses on building safe, sustainable, and real-time Cyber-Physical Systems (CPS). Modern CPSs increasingly rely on ML components for perception and control, making safety guarantees harder to establish while also becoming highly resource-intensive. My work addresses this by managing and provisioning computational resources in ML-enabled CPS to minimize resource usage while satisfying all safety criteria and real-time performance requirements. I co-design controllers and schedulers with formal safety and timing guarantees for CPS with DNN-based perception on edge-cloud platforms, and apply learning-based techniques to formal verification and synthesis problems in safety-critical CPS. My work has applications in automotive, robotic, and industrial control systems.

Education

• Advised by Prof. Samarjit Chakraborty.
• Promoted to Ph.D. Candidate in Fall 2025.

• Ranked 1st out of cohort of 40 students in every semester.

Publications

See Publications.

Projects and Collaboration

• May 2025 – Aug 2025, University of North Carolina at Chapel Hill: Summer Research Collaboration with General Motors. Timing analysis of the 10Base-T1S Ethernet protocol for automotive networking, under Prof. Samarjit Chakraborty, in collaboration with General Motors.

• Nov 2023 – Jun 2024, University of Calcutta: Undergraduate Thesis. Drone and Cab-based Hybrid Package Delivery Optimization, under Prof. Sunirmal Khatua. Proposed greedy algorithms for last-mile package delivery optimization using third-party cabs for warehouse-to-pickup-point transport and drones for final delivery.

• Oct 2023 – Jan 2024, Indian Statistical Institute: Research Intern. BMC Engine Sequencing with GNN Embeddings of Hardware Circuits, under Prof. Ansuman Banerjee. Researched the selection of Bounded Model Checkers for the verification of safety properties of circuits from the HWMCC competition.

• Jan 2024 – Mar 2024, Starinco: Machine Learning Developer Intern. Worked in a team of 7 to train computer vision models for automation and auditing in industrial and manufacturing settings.

• Oct 2023 – Dec 2023, Indian Statistical Institute: Research Intern. Verilog Code Generation and Debugging using LLMs, under Prof. Ansuman Banerjee. Researched the applications of Large Language Model (LLM) based tools for Electronic Design Automation and Formal Verification.

• Jun 2021 – Aug 2021, Google Summer of Code at Debian: Software Developer Intern. Wrote build scripts and set up a CI/CD pipeline to compile and update the Android SDK tools package for Debian. Worked in a team of 5 people. See: wiki.debian.org/PrateekGanguli/SummerOfCode2021

Talks

• “Scheduling Control Tasks Using Safety-Guided RL,” in International Conference on Applied Algorithms (ICAA), 2026.
• “Reducing Cloud Communication Bandwidth in Edge-Cloud Partitioned Neural Networks for CPS,” in International Conference on Computational Technologies and Electronics (ICCTE), 2025.
• “Repairing Control Safety Violations via Scheduler Patch Synthesis,” in International Conference on Cyber-Physical Systems (ICCPS), 2025.

Service to the Community

• Repeatability Evaluator, International Conference on Cyber-Physical Systems (ICCPS), 2026.
• Reviewer, International Conference on Robotics and Automation (ICRA), 2026.

Teaching Experience

• Teaching Assistant, UNC Chapel Hill, Programming Intelligent Physical Systems, Spring 2026.
• Student Mentor, IIT Madras, Machine Learning Practice, Linux System Commands, Summer 2023.
• Teaching Assistant, Uni. of Calcutta, Data Structures and Algorithms Lab, Fall 2022.