Building privacy-first,
on-device AI for healthcare.

Model Evaluation

We Tried a Bigger Model. Here's Why We Shelved It.

A health-specialized model looked like an obvious upgrade. It took a missing off-ramp, two false leads, and a rigged test before we got an honest answer about whether it belonged in the app.

16–19 min Intermediate / Advanced 16 Jul 2026

Performance Engineering

The Problem Was Never the GPU

A real performance investigation that started with the obvious suspect, the inference backend, found a lifecycle bug one layer up, then found a second reload nobody had questioned, and put real device numbers behind GPU and NPU.

18–21 min Intermediate / Advanced 13 Jul 2026

AI Architecture

LLMs vs SLMs: Why Bigger Isn't Always Better

The industry is obsessed with parameter counts. But model size is only one dimension of quality, the real engineering challenge is choosing the right model for the task.

12–15 min General Audience 9 Jul 2026

Security Architecture

Every System Needs a Deliberate Exit Door

We spend enormous effort protecting how data enters a system, and almost none designing how it leaves. Why outbound paths are the trust boundaries that matter.

13–16 min Intermediate 8 Jul 2026

Edge AI

Democratizing AI

How Small Language Models and Edge AI could make artificial intelligence accessible to billions of people.

12–15 min General Audience 8 Jul 2026

Architecture

The Engine Isn't the Car

Why using an external AI model doesn't make FarmakoMed beholden to its vendor, and why sovereignty lives in the architecture, not the components.

14–17 min Intermediate 5 Jul 2026

Architecture

AI Should Wait for the User, Not the Other Way Around

Why building seamless on-device AI forced us to redesign our application architecture instead of our machine learning models.

13–16 min Intermediate 4 Jul 2026

AI Economics

The Hidden Cost of AI Isn't Compute. It's Dependency.

How tokens quietly became the currency of artificial intelligence, and why on-device AI offers a different vision.

12–15 min General Audience 3 Jul 2026

Privacy by Architecture

Privacy by Architecture

Why local AI changes software architecture, not just privacy policies.

8–10 min Intermediate 30 Jun 2026

Local AI

Choosing a Model for On-Device Healthcare AI

Why selecting the "best" open model matters less than understanding your product's constraints.

10–12 min Intermediate 2 Jul 2026

Diagnostics

Diagnostics Without Telemetry

Debugging on-device AI when you can't rely on server logs or remote sessions.

14–17 min Intermediate / Advanced 21 Jul 2026

Performance

Battery-Aware AI Inference on Mobile

Why battery life should be treated as a first-class feature, not an afterthought, when building AI-powered mobile applications.

10–12 min Intermediate 7 Jul 2026

AI & Sustainability

Can AI Be Sustainable?

Why the future of artificial intelligence should be measured not only by intelligence, but also by its environmental footprint.

12–15 min General Audience 7 Jul 2026

Mobile Engineering

Building Self-Healing Android Applications

Why keeping an app alive should never make it slower, treating a long background stay as a health check instead of waiting for a restart.

13–16 min Intermediate 7 Jul 2026