About | Logan Thorneloe

About | Logan Thorneloe I am an ML infrastructure engineer at Google Ads, partnering with researchers to design high-performance workflows, optimize fleet-wide efficiency, and develop agentic tools to improve ML experimentation velocity. I am also the author of AI for Software Engineers, a publication that gets software engineers hands-on with AI and is read by over 13,000 developers. You can find me on X and Substack, or chat by shooting me an email. What I'm working on - • Co-developing an MLE Agent with Google DeepMind. - • Improving the bulk inference user experience for Gemini in Ads. - • Publishing AI for Software Engineers(see above). - • Creating a foundational guide to AI for Software Engineers. Previous Experience - • ML Infrastructure Engineer at Google, building developer AI tools and improving machine learning velocity. - • ML Infrastructure Engineer at Microsoft, connecting new compute to training runtimes and managing asset reuse across the fleet. - • Software Engineering Intern at Microsoft Security Response Center, developing open-source hardware monitoring systems in Rust. - • Research Assistant at BYU, developing CNN-based MRI semantic segmentation models. What I Believe - • Machine learning is about creating excellent products and services people can use and apply. The key to building a great machine learning system is to be a great engineer. - • The future of AI belongs to engineers. The impact of AI research relies on how well we can usefully put it into the hands of users. - • I don't understand that which I can't build. Building is the fastest way to turn ideas into real understanding. - • You can just do things, but don't do too much. It's better to take a step back and ensure you're headed in the right direction. - • Always be yourself. In work, life, and relationships, everything tends to work out better when you do. Also, everything's more fun with a bit of whimsy. - • Ask for help. The best way to get better at somethin