pepe:curious:my_corner

πŸ‘‹ hey, i'm lovepreet singh

mle @cred | youtube β€’ x β€’ linkedin β€’ dev β€’ github

i’m an applied ml + systems engineer building bottom-up β€” from os schedulers and drivers to cuda experiments and llm pipelines.

my goal is to contribute to ai-first infra teams where deep fundamentals meet product impact. i like working close to metal, math, and users.


dissecting linux schedulers β€” implementing a toy cfs simulator

cover tags: os, scheduling, c, simulation
a clean breakdown of how linux’s cfs works, with a custom simulator demonstrating vruntime, run queues, fairness, and weighting.


πŸ“š all posts

k-means

adapted k-means for color segregation

python β€’ computer vision β€’ ml

optimizing centroid initialization to quantify plant health from raw pixels β€’ Supported by NNSF China govt.

k-means

building soil moisture monitors with arduino + android

arduino β€’ python β€’ c

integrating sensors, microcontrollers, and mobile apps into geotechnical engineering labs

usb driver

usb driver in c β€” writing a minimal driver from scratch

kernel β€’ drivers β€’ c

a minimal usb driver, understanding interrupts, urbs, descriptors, and driver–kernel communication.

chatroom c++

a multithreaded chatroom in c++ (boost.asio)

networking β€’ c++ β€’ concurrency

a chatroom server built using boost.asio and threads β€” a practical intro to async network programming.

p2p java

peer-to-peer file sharing in java + nginx + docker

dist-sys β€’ java β€’ infra

a p2p file-sharing system with multi-client coordination, reverse proxy, containers, and cloud.

linux cfs

dissecting linux schedulers β€” toy cfs simulator

os β€’ scheduling β€’ c

a hands-on deep dive into linux's cfs with a custom scheduler simulator.


🧭 what this blog is

πŸŒ‰ why you care

my work sits at the intersection of systems, ml, infra, and optimization.
if you're working on ai infra, model training systems, inference pipelines, or low-latency distributed compute, this blog reflects exactly how i think and build.


🐸 about pepe:curious

a small corner of the internet where i document experiments, failures, insights, and implementation notes.