Chris Lovett

关于

I’m a Principal Software Engineering Manager who joined Microsoft in ’97 and I’ve worked on everything from XML, IE, .NET, SQL, Visual Studio, and Windows and now in Microsoft Research.

I have done several key incubation projects over my career at Microsoft that were close partnerships with Microsoft Research. The first was an experiment in data integration into mainstream programming languages inventing a new language named X#.  Many of the ideas ended up in C# LINQ. The second was an exploration into graphical visualization of code dependencies collaborating with MSR which led to the DGML diagrams and the Visual Studio Code Map features.  The other was a secret project in the AI consumer device space that also came out of Microsoft Research.

I enjoy working with Microsoft Research so much I officially joined Microsoft Research in 2015 to research safe and robust cyber physical systems using the P programming language.  I have also worked on AirSim (opens in new tab) and the ELL (opens in new tab) Neural Network Compiler for embedded devices and the Coyote (opens in new tab) project for systematic testing of asynchronous code. I’ve also worked on the Lean Programming Language (opens in new tab) and Neural Architecture Search (opens in new tab) and most recently the use of AI in Game Testing.

I have over 25 experience years working on everything from IBM, OS/2, Linux, Apple, Android, and Windows platforms and I enjoy tinkering with embedded systems using ARM Cortex-M class chips, my favorite being those from Nordic Semiconductor. I have way too much experience with C/C++ & C#, quite a bit of Javascript & Typescript & nodejs, and about 8 years with Python since many research projects use Python, especially pytorch in the AI space.  Funny thing is when researchers first asked me to write some python I took a look and was repulsed by the required indentation, and no curly brackets, but I have to say the language has grown on me over time.  It is now my go-to language for building quick little tools.  I think the main secret to it’s success are the amazing libraries from numpy, opencv, and pytorch.  Note that all of those are implemented in C++.  So python is really just a glue language, and a very slow one at that.  But the libraries are amazing.