Ajeet Kumar

Ajeet Kumar

Applied Mathematician with Specialization in Full Stack Machine Learning and Data Science

System Design Part 9

less than 1 minute read

Published:

System Design from First Principles - The Mathematics of Attempts

Every computer system core feature is the failure, therefore we needed to make system to work when everything is perfect to world where everything is broken.

You May Also Enjoy

System Design Part 8

less than 1 minute read

Published:

System Design from First Principles - CAP Theorem & PACELC

System Design Part 7

less than 1 minute read

Published:

System Design from First Principles - Sharding: When One Disk is Not Enough

Operational Ecosystems of LLMs in Python - Part 2

1 minute read

Published:

Module 2 : Re-Engineering and Verification Workflow

Module‑1 showed you that LLMs are not magic but a composition of code, math, and hardware, all reachable from Python if we know where to look. It traced how LLMs sit in the stack: Python as glue, PyTorch as the modeling layer, and hardware‑specific backends (CPU, GPU, etc.) as the execution engines.

Operational Ecosystems of LLMs in Python - Part 1

7 minute read

Published:

The first question came into the my mind when i started thinking about the operational ecosystem of LLM is that Why LLMs matter in the AI ecosystem ? LLMs are a core component of modern artificial intelligence. In practice, we use Python extensively for almost any task in AI and data‑driven decision‑making from simple scripts to complex production pipelines. Today, the ecosystem surrounding Python is rich and specialized for nearly every problem domain, there exist dedicated tools that can be easily integrated because they live inside Python’s operational environment. So, they are the natural evolution of this ecosystem they turn language into a computational interface but still sit on top of same stack of tools, hardware and orchastration layers that power the rest of AI.