Notes from the build
Learning out loud.
Engineering decisions, experiments, and lessons worth sharing. A few things I’ve learned while making things work.
Why iOS Deferred Deep Linking Was a Pain Compared to Android
Android had an official install referrer. iOS had limitations, unreliable device matching, and a lot more work. Here is how I ended up with an IP-based referral workaround that is useful in production, even with its limitations.
Quick Health Checks After React2Shell: What Every React/Next.js Developer Should Know
Recently, many React/Next.js projects were affected by the React2Shell remote code execution, a critical score 10 vulnerability. Even if your app appears normal, the underlying machine may still show signs of compromise. Here are a few quick checks developers can run to get an initial sense of system health.
Building flutter_native_ocr: A Native OCR Flutter Plugin from Idea to Production
In this post, I'll share my journey of creating flutter_native_ocr, a Flutter plugin that brings native OCR capabilities to mobile apps. This plugin leverages Apple Vision framework on iOS and Google ML Kit on Android to provide accurate, on-device text recognition without requiring internet connectivity.
The Evolution of Coding: From Machine Coding to Modern Languages to LLM
The landscape of programming has undergone tremendous evolution over the years, moving from low-level machine languages to high-level languages that abstract complex operations. This evolution has made coding more accessible and simpler, enabling developers to focus more on solving problems than on managing intricate details of the hardware. Here’s a brief overview of this journey through different programming languages, showcasing a simple "Hello, World!" example in each.
How I used on-device ML and OpenAI to reduce a service cost by 40x
In this post, I'll share how I used on-device machine learning and OpenAI to reduce a service cost by 40x. This project involved optimizing a service that relied on multiple third-party API for processing user requests. By leveraging on-device machine learning and OpenAI's GPT 4o-mini model, I was able to significantly reduce the service's dependency on the external API, resulting in substantial cost savings.
Prefer a quieter way to keep up? Subscribe via RSS