Build logs, architecture notes, and product decisions.
Every post is loaded from markdown in content/blog. Keep adding your .md files there and this section will render automatically.
Why I Chose Gemma 4 E2B for Subra AI: The Reality of Running LLMs on Mid-Range Phones
A practical look at why Gemma 4 E2B is a strong fit for running a private, multimodal AI assistant on mid-range phones.
React Native + Gemma: Building a Fully Offline AI Hydration Coach
How I embedded an on-device Gemma model in React Native for a sovereign, offline hydration coach without cloud APIs or token bills.
Stop Shipping 20 Locale Files in React Native: On-Device Translation for Dynamic Language Packs
Learn how to avoid shipping dozens of locale JSON files by generating and caching a dynamic language pack in React Native.
How to build a Liveness-Check Login with face-api.js and React
Build a login that requires a liveness check (e.g., 'Blink twice to login') to prevent photo-based bypasses — using face-api.js in the browser.
The CocoaPods Sunset: What Dec 2, 2026 Means for Your React Native App
Explore what the CocoaPods read-only transition on December 2, 2026, means for React Native apps, the shift to Swift Package Manager (SPM), and how to prepare your projects.
The Right Tool for the Job: A Guide to AI Models by Data Type
Understand the different types of AI models and their ideal use cases, from text to vision to audio.
Common LLM Parameters To Know For RAG
A practical, provider-neutral guide to the key retrieval and generation parameters that control RAG quality, latency, and cost.