Hi, I’m Sai Saketh Cherukuri, a final year M.Sc. student in Mathematics with a specialization in Computer Science at the Sri Sathya Sai Institute of Higher Learning. My interests lie at the intersection of machine learning and high performance computing, with a strong focus on model efficiency and system level optimization. I work extensively with GPU acceleration and performance tuning, aiming to build solutions that are both accurate and computationally efficient. My recent projects include benchmarking models using the MLPerf inference suite to evaluate and optimize throughput, as well as developing a periocular based age and gender recognition system using PyTorch and modern deep learning techniques. These experiences reflect my interest in combining mathematical rigor with practical implementation to build scalable, high performance ML systems.
Selected Work
A collection of my project spanning work across Machine learning, Deep Learning, LLM fine tuning, Inference optimization, GPU Profiling, tooling, and low level programming.
Built a reproducible benchmarking harness for MLPerf Llama 3.1 8B inference on a single NVIDIA H200, studying throughput, accuracy, and tuning tradeoffs across scheduling, quantization, and memory settings.
Developed a periocular recognition model for age and gender prediction, combining image preprocessing, model training, and evaluation workflows in PyTorch.
Created a static analysis toolkit for C++, CUDA, and Python with a JavaFX interface as part of my M.Sc course requirement.
Implemented a compact two pass SIC/XE assembler in C++ that builds a symbol table, resolves directives, and emits standard object program records.
Built a recursive descent parser in C for a simplified Pascal like language, validating declarations, assignments, expressions, and basic control flow syntax from token streams.
You can reach me directly at ssaketh.2002@gmail.com : Feel free to reach out!