Architecture decision that paid for itself immediately:
Swapped the ML model in my Android app from logistic regression to a neural network.
Python script: 7 changes. Android code: 0 changes.
How interface design makes model upgrades free:
https://t.co/6zLAha1XKe
@towards_AI
I just published my first book: "Machine Learning for Android Engineers."
It teaches you how to train a model in Python, convert it to TFLite, and run it on-device in an Android app. The final model is 1.8 KB.
Here's what I learned building it.
The book covers the full pipeline:
Python training + TFLite conversion Android integration (Clean Architecture + Hilt + Compose) Production hardening (testing, crash reporting, kill switch)
134 pages. No fluff. Every line of code explained.
๐งฑ Breaking the Monolith: A Practical, Step-by-Step Guide to Modularizing Your Android AppโโโPartย 2
{ by Vortana Say } from @hashnode https://t.co/G8ltTEpC4r
๐ Welcome Kotlin Multiplatform in @JetBrains_Fleet ๐
The new tooling allows you to streamline the development of multiplatform apps in a single IDE, making the process simple and enjoyable. Features include:
๐ Simplicity: Auto-generated Run Configurations for Android and/or iOS.
๐ Polyglot programming: No need to switch to a different editor when working with the native code in your multiplatform project.
๐ Swift-Kotlin cross-language support: Fleet provides cross-language ัode navigation, refactoring, and debugging.
Give it a try, and share your feedback!
https://t.co/tXAXBOjsq7
๐ I just finished Day 47 of the #100DaysOfSwiftUI at https://t.co/upOngnUbxs via @twostraws
Basic Habit Tracker:
- List of habit
- Form add new habit
- Detail screen
- Add / Plus button to increase / decrease completion count
- Load and save data