My path here was unconventional—I was originally a pre-med student and was accepted into a T5 medical school. However, during that journey, I realized that my impact could extend far beyond the clinic. Through machine learning and technology, I could help optimize the inefficiencies that plague healthcare systems, develop diagnostic tools that scale to millions of patients, and build solutions that physicians could never accomplish alone. This realization led me to transition into computer engineering, where I now combine my understanding of medicine with my passion for AI to create technology that truly makes a difference.
I am currently pursuing my Master's degree in Computer Engineering at NYU Tandon School of Engineering. My academic journey—spanning both medicine and engineering—has provided me with a unique perspective on solving complex problems, along with a strong foundation in programming, ML systems, and hardware-software integration.
I am passionate about optimizing machine learning systems at the hardware level and exploring the intersection of AI and robotics. My research interests include:
- Hardware Optimization: Improving performance and efficiency of ML workloads through kernel optimization, memory management, and custom accelerator designs
- Reinforcement Learning for Robotics: Developing intelligent agents that learn to interact with and manipulate the physical world through trial and experience
- High-Performance Computing: Leveraging GPU architectures and parallel computing paradigms to accelerate training and inference
- ML Systems: Building scalable, efficient infrastructure for deploying machine learning models in production
Throughout my academic and project work, I have developed expertise in:
- GPU Programming: CUDA, Triton, OpenCL
- ML Frameworks: PyTorch, JAX, Pallas, TensorFlow
- Next-Gen AI Tools: Mojo, Modular MAX
- Programming Languages: Python, C/C++, Golang, Rust
- Deep Learning: Neural Networks, Transformers, Reinforcement Learning
- Systems: Linux, Docker, Distributed Computing
If you'd like to connect or learn more about my work, please reach out!
GitHub: koshiq
LinkedIn: koshiq