sahilnyk

Simple explanation of why ChatGPT-like models cost so much to run and how companies make them faster.

The time I processed the same payment twice, killed my queue with one bad message, and burned $200 on polling.

Simple breakdown of when serverless makes sense and when it's just extra cost.

AI Code Quality Guardian - Monitor how AI changes your code, track requirement alignment, and catch vulnerabilities. Acts as a production-grade quality gate for AI-assisted development, analyzing changes, running tests, and keeping AI on track.

A FastAPI scaffolding CLI tool that generates production-ready project structures with zero dependencies. Supports JWT auth, multiple databases (PostgreSQL/MySQL/SQLite), Docker, and Terraform infrastructure setup.

Full Stack Engineer

Mar 9 - Present

@Eldridge Morgan

Building enterprise solutions that actually scale. Architected multiple internal tools from scratch, managed infrastructure with Terraform, and shipped features that handle real production load. Basically turning coffee into code and keeping the servers happy.

Machine Learning Intern

Jul 15 - Aug 15

@CTTC Bhubaneswar

Orchestrated the entire ML pipeline for a computer vision recognition system. Wrangled datasets, trained models, and deployed them to actually work in the real world. Turned messy data into something that could see and understand images.

Software Engineer Intern

Apr 5 - Jul 12

@Infoware System

First real dev gig energy. Built features, squashed bugs, and learned what production code actually means. Collaborated with the team on client projects, wrote APIs, fixed frontend issues, and shipped code that people actually used.