Developed resource management mechanisms to improve GPU utilization and efficiency for deep learning training workloads as part of the Resource-Aware Data Systems team, under the supervision of Prof. Pınar Tözün.
Thesis: "Collocation and Memory Estimation for Efficient GPU Resource Management in Deep Learning Training"
★ ★ ★
Designed and evaluated an energy-efficient on-chip memory design for GPUs with negligible performance overhead, under the supervision of Prof. Hamid Sarbazi-Azad; leveraged GPGPU-Sim and developed C and Python code for evaluation.
Thesis: "Unifying L1 Data Cache and Shared Memory in GPUs"
GPA: 17.20/ 20
★ ★ ★
Built a solid foundation in computer science through rigorous coursework and extensive self-study, with particular interest in programming, logic and digital circuit design, computer architecture, and Verilog. Strengthened practical skills in MATLAB and C++ through a capstone project on vehicle distance estimation using laser and computer-vision methods.
Project: "Car Distance Estimation Using Laser and Computer Vision Methods"
GPA: 18.42/ 20