Siddesh Sambasivam Suseela

Siddesh Sambasivam Suseela

Full Stack Software Engineer

Hypotenuse AI

Hey there πŸ‘‹

I’m Siddesh and this is my digital commonplace book.

Explore More

Interests
  • Software Engineering
  • Natural Language Processing (NLP)
  • Deep Learning
  • UI/UX Design
  • Investments
  • Physics-informed Knowledge Discovery
Education
  • Bachelor's in Electrical and Electronic Engineering, 2018-2022

    Nanyang Technological University

Recent notes

BitNet: Scaling 1-bit Transformers for LLMs

Background LLMs is just a bunch of optimized matrices in N layers with floating-point weights. Generation of LLMs means doing a bunch of matrix multiplication. And doing floating point operations is expensive in terms of energy consumption.