Blog

Writing

Technical notes on ML engineering, data infrastructure, and the realities of deploying AI in production.

Latest

What an AI strategy actually looks like (hint: it’s not a deck)

Most AI strategies are 80 slides of buzzwords that never lead to production. Here’s what a useful one contains: specific use cases, data readiness assessments, and honest timelines.

Feb 10, 2025 12 min
Jan 28, 2025 8 minGovernance

EU AI Act: practical guide for Turkish companies

If you sell to European markets, the EU AI Act affects you. Here’s a no-nonsense breakdown of what it means for your AI systems and what to do about it.

Jan 15, 2025 15 minEngineering

Getting ML models to production: what actually works

90% of ML models never make it to production. We share the patterns that work for us: simple serving, automated retraining, monitoring that catches drift before it costs you.

Jan 3, 2025 10 minInsights

Where GenAI works in enterprise (and where it doesn’t)

LLMs are powerful but not a silver bullet. Five use cases where we’ve seen real ROI, and three where companies keep wasting money.

Dec 20, 2024 7 minData

Your AI problem is probably a data problem

Before investing in models, invest in data. Our framework for assessing data quality and what to fix first when everything looks messy.