High-performance computing (HPC) with NPUs and GPUs to enhance AI capabilities.

• Engineered an optimized parallel library, leveraging custom-designed GPU kernels in SYCL for advanced numerical differentiation, aimed at enhancing computational analysis efficiency.• Significantly enhanced loop closure efficiency in real-time robotics (SLAM) applications, achieving a 400% increase in processing speed.• Dhruv Kumar, Shishir Gopinath, Karthik Dantu, Steven Y. Ko. JacobiGPU: GPU-Accelerated Numerical Differentiation for Loop Closure in Visual SLAM. (ICRA 2024) Yokohama, Japan.

• Managed a residential community of 100 graduate students, enhancing their academic and personal well-being through support, conflict resolution, and diverse events, while fostering a positive living environment and upholding university policies.

• Led and implemented diverse events in a residence hall, fostering community engagement through culturally enriching programs, effective collaboration with staff and vendors, and strong communication skills, resulting in high participation and a vibrant community atmosphere.

CMPT 300: Operating Systems I [Spring 2023]CMPT 218: Systems Programming [Fall 2024]◦ Designed and implemented comprehensive unit test scripts using the Google Test framework, seamlessly integrating them with Continuous Integration (CI) via GitHub Actions for automated grading and code quality assurance.◦ Held review sessions, office hours and one-on-one appointments.

• Developed a serverless inventory management backend using Node.js on AWS Lambda for dynamic scalability. Integrated with MongoDB Atlas for effective multi-location tracking and synchronization across 10 gas stations.• Leveraged MindsDB for predictive analytics in inventory management. Utilized artificial intelligence and machine learning for analyzing sales data and trained models to forecast sales trends, significantly enhancing inventory decisions.• Optimized inventory management using MongoDB Change Streams for real-time updates, enhancing responsiveness and reducing stock discrepancies in high-volume transaction settings.• Achieved a 50-hour weekly reduction in manual inventory management across all stations, streamlining operations and yielding significant cost savings through tech integration and data-driven methods.

CSE331: Introduction to Algorithm Analysis and Design [Fall 2017, Fall 2018]CSE 250: Data Structures [Fall 2018]CSE 191: Discrete Structures [Spring 2018]CSE 199: Internet of Things [Fall 2018]◦ Developed a grading tool to evaluate programming assignments written in Java, Python, and C++.◦ Made weekly walkthrough videos for theoretical assignments.◦ Taught two one hour recitation/lab sessions to ~40 students.◦ Held one-on-one appointments to help students with various problems.◦ Graded assignments/exams.

• Formulated a multi-threaded benchmark suite for analyzing garbage collector behavior in various functional languages, including Haskell, Scala, and others, employing diverse memory exhaustion strategies to simulate intensive workloads.• Publication: Jeffrey C Murphy, Bhargav Shivkumar, Amy Pritchard, Grant Iraci, Dhruv Kumar, Sun Hyoung Kim, and Lukasz Ziarek. A Survey of Real-time Capabilities in Functional Languages and Compilers. Concurrency and Computation: Practice and Experience. 2019; 31:e4902. https://doi.org/10.1002/cpe.490

- Develop social and educational events for residents.- Hold monthly interactive sessions with all the residents.- Lead all the residents and respond to all emergency situations.- Work to build community within the dormitory.- Work with students and all the administrators of campus living as a team.

- Worked in a professional office which entailed check-ins and check-outs of the guests, overlooking the mailing system, resolving any issues that guests might have and much more.- Worked to build community within the dormitory.- Responded to emergency situations.- Enforced rules of the residence halls.- Acted as a liaison between students and administrators.- Developed social and academic events for conference guests.- Held Monthly Interactive sessions with all the conference guests.