This repository contains my notes and code related to MLOps and MLSys. It serves as a personal study resource for understanding the concepts, tools, and best practices in the field of machine learning operations and systems.

MLOps and MLSys

What is the difference between MLOps and MLSys?

MLOps (Machine Learning Operations) focuses on the practices and tools for deploying, monitoring, and maintaining machine learning models in production. It encompasses the entire lifecycle of machine learning models, from development to deployment and ongoing management. MLSys (Machine Learning Systems) focuses on the design and implementation of the underlying systems and infrastructure that support machine learning workloads. This includes hardware, software, and frameworks that enable efficient training and deployment of machine learning models.

There is heavy overlap between the two fields, as both MLOps and MLSys are concerned with the practical aspects of machine learning in production environments. However, MLOps is more focused on the operational aspects of machine learning, while MLSys is more focused on the technical and infrastructural aspects.

Contents

• ML-raw - Raw code and notes related to machine learning models and algorithms.
• MLOps
  • 2. Chip Nyugens - form a roadmap for learning MLOps - 10 course modules on MLOps and MLSys
  • 1. Lil'Log - on Agentic AI assistant systems
  • 3. deeplearnings .ai course
  • 4. MS github modules
• MLSys Zero to hero in LLMs - https://karpathy.ai/zero-to-hero.html CUDA understanding - https://modal.com/gpu-glossary/device-hardware/cuda-device-architecture

inPipe: Course work on LLMs basics - - LLM while inference - in C++ - Fast.ai course -> - https://training-large-models-course.github.io/ -> GPT in basic C++ - Youtube courses on LLMs