River Riddle Email & Phone Number

Santa Clara, Ca, Us

Santa Clara, Ca, Us

Everywhere, Us

I lead a team of engineers that contribute tools and infrastructure for the Mojo programming language.With the team, I drive designs, roadmaps, and am heavily involved in the implementation of a large surface area of the Mojo programming language:* The Mojo build system and package management infrastructure.* The Mojo compilation model, building packages, executables, JIT infrastructure, incremental/cached compilation, and much more.* An LLDB-based Debugger (without forking LLVM or LLDB!)* The generation and validation of Mojo API documentation.* The Mojo Language server and VS Code extension, supporting a wide range of modern language features (including jupyter notebooks!).* A black-based formatter with support for Mojo.* Support for building and developing Mojo on every major platform: Linux, Mac, Windows.* An LLDB-based REPL and Jupyter notebook for Mojo.* A unit and documentation testing framework built for Mojo.There's really too much that the team does, but it's just about everything involved with developing, build, or interacting with the Mojo programming language!For a large period of time, the team also directly managed the release process for shipping the Mojo SDK. This includes the development for nightly builds, uncovering and resolving user platform issues, and much more!

Everywhere, Us

Joined Modular as the first member of the team after the two founders.* Continued to maintain and evolve the MLIR ecosystem, including the introduction of an MLIR bytecode format, "resources", PDL(L), the death of the "standard dialect", and much much more.* Managed the effort for productionizing and shipping the Mojo Language publicly * Led a team of 3 other engineers, defined roadmaps, tracked work, handled team coordination, etc. * Built the Mojo packaging and module system * Developed an LLDB-based debugger, REPL and Jupyter notebook (with native matplotlib support!) * Managed Mojo's JIT infrastructure (based on LLVM-Orc) * Drove support for Mojo on each major platform: Linux, Mac, Windows * A lot lot more, but see the follow-on explicit Manager role for some more information!

Mountain View, Ca, Us

I was a main Architect of MLIR Compiler Infrastructure. I was responsible for the design and implementation of most of the core pieces of MLIR; this includes the IR structure, tooling, documentation (including the tutorial and a large portion of mlir.llvm.org), and much much more. MLIR infrastructure is now used pervasively within Google for Machine Learning code generation. Aside from Machine Learning, I evangelized the use MLIR on many other computational use cases inside of Google (of which I can't really talk about).

Mountain View, Ca, Us

I led various code generation and optimization efforts on a custom VLIW LLVM backend, for internal hardware (that I can't talk about much).

San Mateo, California, Us

I led various key areas of compiler technology, including designing novel algorithms and infrastructure for code size optimization for small targets (including outlining), targeted size and runtime optimization for PlayStation platforms (some of which I can't talk about), compiler tooling (for optimization visualization). I also often visited PlayStation Studios game teams to provide direct assistance with compiler optimization and feature design.

San Mateo, California, Us

Designed various PGO based optimization algorithms, including partial inlining and function outlining, re-implemented various PS4 SDK tools for efficiency and cross platform compatibility. Revamped and analyzed performance metrics for internal game benchmarks.