Daniel Ellis - CEO, Founder at Tiny Little Keys

Daniel Ellis

CEO, Founder at Tiny Little Keys

United States

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Daniel Ellis is a founder and seasoned engineering leader with 14 years building scalable ML systems, compilers, and backend infrastructure, most recently founding Tiny Little Keys after senior roles on TensorFlow and PyTorch-on-IREE at Google. He combines hands-on ML compiler and runtime experience (IREE, torch-mlir) with deep production engineering chops from work on reddit’s backend and infrastructure, including distributed systems, deployment, and metrics collection. His open-source contributions span high-impact projects—helping bridge PyTorch to MLIR and improving release automation and tracing—demonstrating a knack for making research-grade tooling production-ready. At Reddit he touched core relevance and ML infrastructure, and earlier work on Diamond and memcached collectors shows attention to robustness and observability. Comfortable switching between low-level C runtime work and high-level MLOps automation, he prioritizes practical solutions that scale. Based in the United States, he blends research-facing compiler expertise with pragmatic backend and DevOps instincts that few engineers straddle so effectively.

14 years of coding experience

14 years of employment as a software developer

Boston University

Github Skills (30)

socket-io10

pytorch10

github-ci10

runtimes10

back-end-development10

python10

cassandra10

testing10

machine-learning10

mlr10

compiler-compiler10

socket10

memcached10

compiler10

backend10

Programming languages (14)

MDXJavaC++CRustGoHTMLJupyter Notebook

Github contributions (5)

llvm/torch-mlir

Apr 2020 - Dec 2022

The Torch-MLIR project aims to provide first class support from the PyTorch ecosystem to the MLIR ecosystem.

Role in this project:

MLOps Engineer & Backend Developer

Contributions:24 reviews, 90 commits, 28 PRs in 2 years 8 months

Contributions summary:Daniel focused on enhancing the torch-mlir project's support for PyTorch and MLIR ecosystems, primarily by implementing and improving the build process and backend compatibility. This included creating a GitHub Action for pip-compatible releases, which streamlines the package distribution process, and addressing tracing issues related to dictionary inputs, improving the user experience. Furthermore, the user added support for mv decomposition, enhancing the project's ability to handle matrix-vector operations.

pytorchcompilermlir

reddit-archive/reddit

Jan 2016 - Oct 2016

historical code from reddit.com

Role in this project:

Backend & DevOps Engineer

Contributions:21 commits, 6 PRs, 21 pushes in 9 months

Contributions summary:Daniel primarily contributed to the backend infrastructure and deployment aspects of the reddit project. Their work included enabling and configuring the RabbitMQ management interface, adding gzip support for event collection, and fixing issues related to Zookeeper. They also implemented and refined Cassandra models for IP event tracking, incorporating TTL and reversing result functionalities. Furthermore, the user made modifications to the OAuth2 implementation, adding statistical tracking and introducing logic to handle mobile app gilding events.

pythonprawjavascripthistoricalarchiver

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.

Request Free Trial