Bio: 19yo. Undergraduate student in Data Platform Laboratory, the Faculty of Environment and Information Studies at Keio University, majoring in Computer Science.
Skills / Interests: Common Lisp, Deep Learning, DL Compilers (e.g., Polyhedral Compiler, Auto Scheduler), Neural Network Quantization (especially in CV models), Accelerating scientific computations used in deep learning (e.g.: SIMD, GPU Programming, Parallel/Distributed or Approximated computation), etc. — I am particularly passionate about approaching deep learning and LLM technologies from a low-level perspective and optimizing them.
Projects and Activities:
-
��️ Caten is an experimental deep learning compiler project developed entirely from scratch in Common Lisp. The most challenging aspect of this project is its use of RISC-style instructions, designed to simplify porting to different hardware. Internally, it employs a polyhedral compiler as a low-level IR to optimize auto-generated kernels.
Currently, I am seeking sponsors to continue development after starting university; strong financial support would enable me to further simplify the codebase while expanding support to a wider range of models, hardware, and quantization.
-
I love contributing to open-source tensor compiler projects (e.g., tinygrad).
-
or doing some trivial OSS projects in Common Lisp ...
🏆 Awards and Honors
- U22 Programming Content METI Minister’s award.
- Sōkan no Susume Scholarship, Keio University.
📧 Contact Me: ichndm@gmail.com | yuma.oda@keio.jp
📈 Stats
