James Keane
Software Engineer at Heron Data
United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolJames Keane is a software engineer and AI product builder with ~4 years of experience shipping full‑stack production systems, most recently as a founding engineer at YC W23 startup Unify. He architected and owned an agentic LLM sales automation pipeline that achieved 15%+ conversion to booked calls, rebuilt a fractured developer analytics console into a paid product, and eliminated a fraud vector in a Stripe payments flow within a week. Comfortable across Python, TypeScript, FastAPI, Next.js and cloud infrastructure, he combines backend engineering, frontend UX design, CI/CD, and LLM orchestration to ship end‑to‑end features. He contributes to open‑source ML tooling—helping make Ivy interoperate across PyTorch, JAX, NumPy and TensorFlow—and built property‑based testing to catch cross‑backend edge cases. Outside work he built Realpolitik.world, a real‑time geopolitical dashboard with a 3D globe and an 8‑agent editorial AI pipeline, demonstrating a knack for productized research and creative side projects. Educated in AI at the University of Edinburgh, he pairs startup grit with rigorous technical craftsmanship.
4 years of coding experience
3 years of employment as a software developer
BSc Hons, Artificial Intelligence, BSc Hons, Artificial Intelligence at The University of Edinburgh
Github Skills (14)
pytorch10
machine-learning10
jax10
api10
tensorflow10
python10
converter10
numpy10
transpiler9
translation9
neural-network9
deeplearning-ai9
deep-learning9
transcode9
Programming languages (4)
Github contributions (5)
unifyai/ivy
Mar 2022 - Jan 2023
Convert Machine Learning Code Between Frameworks
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:436 reviews, 70 commits, 217 PRs in 10 months
Contributions summary:James primarily contributed to the development and maintenance of machine learning code conversion functionality. Their work focused on implementing the `ones_like` function, ensuring adherence to the Array API Standard, across multiple frameworks, including PyTorch, Jax, NumPy, and TensorFlow. They also added and modified creation functions with docstrings across different backends, adding unit tests and correcting instance methods. This indicates a focus on the core functionality and compatibility of the project's code translation capabilities.
machine-learningpythontensorflowpytorchnumpy
iamjameskeane/unify
Sep 2024 - Feb 2025
A Python package for LLM dynamic routing through the Unify REST API.
Contributions:21 pushes, 4 branches in 5 months
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.