I write about software engineering. Find my articles here and across leading tech publications.
Have you ever built a brilliant time-series model, one that could forecast sales or predict stock prices, only to watch it fail in the real world? Well, this is a common frustration. Your model works perfectly on your machine, but the moment you deploy it in a Docker container, it seems to develop amnesia. It forgets everything it knew yesterday, making its predictions for tomorrow useless.
Multi-stage Docker builds separate build-time dependencies from runtime requirements, dramatically reducing production image sizes.
Use -slim base images, multi-stage builds, smart layer caching, and chained RUN commands to build lean, fast, and production-ready Docker images.
Dockerfile is an immutable ledger. Look at it this way and optimizing containers becomes intuitive and obvious.
My refreshed personal blog, built with the help of Claude Code.
A simple demonstration of time series prediction using ML model and redis db hosted in Docker container.
A containerized solution for hosting transformer models using Flask, Gunicorn, and Docker with AWS SageMaker deployment support. Build once, run anywhere!
A evolving codebase that demonstrates various techniques to optimize docker image for size and performance.
This project dives into the capabilities of Google's DataGemma LLM and demonstrates how to replicate similar behavior on the Claude LLM through prompt engineering. By combining large language models with real-time data retrieval from Data Commons, we aim to provide accurate and up-to-date responses to statistical queries.