Coffee'n me

TechJul 13, 2026

Building Cognate Local First AI Planner

Cognate is a local-first, privacy-first AI planner built for people who want intelligent scheduling without giving up control of their data. This article explores the architecture behind Cognate, including its deterministic Rust scheduler, offline-first CRDT synchronization, end-to-end encryption, and cross-platform design.

Aditya Tawde

5 min

TechJul 13, 2026

From Side Project to Daily Companion: Building Amadeus AI

After months of experimentation, failures, and countless late-night coding sessions, Amadeus AI has evolved into a production-ready personal AI assistant. This article shares the journey behind building Amadeus AI v6.0.0, the lessons learned, and why I chose a local-first, privacy-focused approach instead of chasing every new AI trend.

Aditya Tawde

5 min

CasualJun 13, 2026

Understanding Coffee Roast Levels — Light, Medium, and Dark

Light, medium, dark — most people pick by bag color. Here's what roast level actually does to your coffee, and what it says about strength.

Aditya Tawde

3 min

CasualJun 13, 2026

Best Coffee Beans for Beginners

The wrong first bean turns people off specialty coffee for good. Here's what to actually look for — and which Indian roasters won't let you down.

Aditya Tawde

2 min

CasualJun 13, 2026

10 Common Coffee Brewing Mistakes and How to Avoid Them

Most bad coffee isn't a bean problem — it's a technique problem. Fix any two of these ten mistakes and your next cup will prove it.

Aditya Tawde

3 min

CasualJun 13, 2026

How to Brew the Perfect Cup

Forget the expensive machine. The perfect home cup comes down to fresh beans, correct water temperature, and one ratio you can't eyeball.

Aditya Tawde

2 min

CasualJun 13, 2026

Arabica vs Robusta — What's Actually the Difference?

Two different species, two completely different cups. Here's what actually separates Arabica from Robusta — and why Robusta isn't the villain you think it is.

Aditya Tawde

2 min

TechJun 13, 2026

Running Open-Source LLMs on a Low-End Laptop: My Experience with Qwen and Gemma

Discover how Qwen 2.5, Qwen 3.5 2B, and Gemma 4 perform on low-end hardware using Q4_K_M quantization. A practical look at local AI, coding assistance, and offline LLM deployment.

Aditya Tawde

3 min