Pat Nadolny

Practical AI engineering insights from a software engineer building with AI.

The AI space is moving incredibly fast and software engineers, myself included, are struggling to keep up. The volume of content is exploding but we can't find the practical guidance we need to build real AI powered software.

I've been building with AI since the early days of the ChatGPT wave and I'm sharing what actually works (and doesn't) in production. From system design patterns to framework choices, deployment strategies to tool evaluations. I focus on the practical challenges of shipping AI software in this fast-moving field.

Having evolved from mechanical to data engineering, through platform and product roles, and now to AI engineering, I understand what it takes to adapt and grow into new disciplines. I'm sharing practical knowledge to hopefully help other engineers make their own transition into AI engineering.

Latest Writing

Published on
April 11, 2026 · 6 min read
Evals: The New Unit Test
ai best-practices ai-engineering
How I think about evals at Cora AI, and why the unit test analogy is the right mental model for engineers coming from a traditional software background.
Read more →
Published on
March 5, 2026 · 12 min read
Building with Voice AI
ai engineering voice-ai
Voice AI has reached a tipping point. Here's how modern voice assistants actually work, what makes them challenging to build, and how to think about choosing your stack.
Read more →
Published on
February 3, 2026 · 4 min read
Throwing Tokens: AI for Speed, Deterministic Code for Certainty
ai
When a product is early, favor AI for speed. Once it stabilizes, pay for certainty with deterministic code.
Read more →
Published on
January 13, 2026 · 4 min read
OpenAI Files as Input: An Alternative to a Custom RAG Pipeline
ai
Exploring how OpenAI's Files as Input feature offers an elegant solution to the RAG vs. context window trade-off, allowing developers to leverage OpenAI's infrastructure instead of building custom RAG pipelines.
Read more →
Published on
December 10, 2025 · 7 min read
What is AI engineering?
ai engineering career best-practices
An exploration of the AI Engineer role, explaining how it combines traditional software engineering with modern LLM tools to build more flexible and powerful systems. Learn what sets this role apart from ML research, data science, and traditional backend engineering.
Read more →
Published on
December 10, 2025 · 2 min read
About this blog
career ai best-practices
Practical lessons from shipping AI software, and why I'm writing them down.
Read more →

Interesting

All notes →

How to turn your AI into a world-class designer
Anshu Chimala, Lenny's Newsletter
LLMs default to the most predictable design choice, which is why AI output looks generic. The fix is pushing the model past that default: give it feeling and intent to work from instead of specs, and treat design as an end-to-end process rather than a single prompt.
21 Lessons From 14 Years at Google
Addy Osmani
Great insights on engineering career growth. Key takeaways: the best engineers are obsessed with user problems, clarity beats cleverness, and your network outlasts every job.

Apps

All apps →

Flect
An AI-powered podcast search engine that finds specific moments in podcasts using natural language queries.
Cut Critic
An AI-driven video editing tool that simplifies editing through plain language.
Job Scouter AI
An AI job scout you train on the criteria that actually matter: comp, culture, commute, unlimited PTO, pets in the office, or anything else.

Subscribe to the newsletter