// FIELD NOTES
Thoughts, postmortems, and notes on machine learning, bioinformatics, and data architecture.
How I designed an end-to-end machine learning pipeline that handles 10k predictions per second using FastAPI, Redis, and Celery.
A practical comparison of REST vs GraphQL from a real project migration — what I gained, what I lost, and what surprised me.
An investigation into how pre-splitting oversampling and feature selection generate artificial clinical accuracy, and if simple rules can compete.
Building an end-to-end data pipeline that scrapes LinkedIn postings, structures dirty descriptions, and visualizes real skill demands in Bangladesh.
Integrating FastAPI, React, and Gemini AI to create a smart campus recovery platform that uses verification quizzes and OTP handovers.
