Evan Jones
Principal ML Engineer at Sony Interactive Entertainment
Raleigh, North Carolina, United States
Summary
🤩
Rockstar🎓
Top SchoolEvan Jones is a Principal ML Engineer with 15 years of experience building internal developer tooling, full-stack platforms, and production ML systems for large tech organizations. He has a track record of conceiving and delivering complex projects that measurably accelerate developer productivity— from Twitter’s self-serve static site platform and reusable React templates to Unit 42’s frontend platform for Fortune 500 clients. His recent work spans generative AI for games and production-grade LLM tooling, including practical contributions to the popular ggml/llama.cpp project to speed prompt loading via session caching. Comfortable moving between frontend UX and low-level backend systems, he combines hands-on implementation skills with program-level leadership and standards-setting. Based in Raleigh, NC, Evan pairs engineering rigor from a BASc in Engineering with a knack for shipping experiments that turn research prototypes into usable products.
15 years of coding experience
15 years of employment as a software developer
BASc Engineering, BASc Engineering at University of Toronto
Github Skills (18)
javascript10
c-language10
session-management10
llama10
llm10
dom-manipulation10
gm10
fileio10
jsdom10
file-handling10
html10
file-processing10
dom-api10
file-access10
cprogramming-language10
Programming languages (9)
Github contributions (5)
ggml-org/llama.cpp
Apr 2023 - Apr 2024
LLM inference in C/C++
Role in this project:
Backend Developer
Backend DeveloperContributions:29 reviews, 23 PRs, 8 pushes in 1 year
Contributions summary:Evan focused on enhancing the functionality of the LLM inference in C/C++ project by implementing a session file format and methods for saving and loading inference sessions. They integrated session saving into the main program, enabling the caching of model states to improve prompt loading times. The changes included modifications to the `main.cpp`, `llama.cpp`, and `common.cpp` files, as well as adding options to save the full output to the session or to cache only the prompt. Furthermore, the user added the feature for grammar-based sampling and fixed prompt cache saving.
cppllm-inferenceggml
jsdom/jsdom
Oct 2010 - Oct 2010
A JavaScript implementation of various web standards, for use with Node.js
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:13 commits in 17 days
Contributions summary:Evan primarily contributed to the JavaScript implementation of web standards within the jsdom repository. Their work focused on improving the handling of HTML entities and text encoding, specifically addressing decoding issues related to HTML entities to ensure proper rendering. The user also made modifications to the DOM manipulation methods, corrected meta tag behavior, and addressed several other related issues. The user demonstrates expertise in Javascript, HTML parsing, and DOM manipulation.
javascriptnodejsjsdomdom-apis
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.