James Kirk
Head Of AI ML at Hue.
Boston, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolJames Kirk is a Head of AI/ML with 11 years of experience building production-grade recommender systems and ML platforms across startups and large tech companies. He blends hands-on engineering (TensorFlow-based recommendation frameworks like tensorrec) with technical leadership roles from Staff ML Engineer at Spotify to VP ML/AI at Hunt Club and CTO/co-founder at Meru. His background spans systems performance, trading infrastructure (contributions to Quantopian’s Zipline and trading_calendars), and robotics, informed by dual master’s degrees in mechanical engineering and business management. Based in Boston, he moves fluidly between research, product, and ops—often improving build and deployment processes while shipping novel algorithms. Colleagues describe him as an engineer who scales both teams and complex personalization systems, with a penchant for integrating sampling strategies and regularization into practical models.
11 years of coding experience
11 years of employment as a software developer
Dublin SchoolMaster's Degree Business Management, Master's Degree Business Management at Worcester Polytechnic Institute
Spanish
Github Skills (28)
recommendation-system10
trading-algorithms10
python10
back-end-development10
finance10
pandas10
machine-learning10
cryptocurrencies10
numpy10
tensorflow10
statistical-models10
cython10
recommender-system10
neural-network10
backend10
Programming languages (2)
Github contributions (5)
jfkirk/tensorrec
Feb 2017 - Feb 2020
A TensorFlow recommendation algorithm and framework in Python.
Role in this project:
ML Engineer
ML EngineerContributions:304 commits, 52 PRs, 235 pushes in 2 years 11 months
Contributions summary:James implemented a novel recommendation algorithm and framework in Python using TensorFlow. Their primary contribution was the development of a ReLURec class, demonstrating an understanding of neural network architectures. The user's work included constructing the computational graph, defining placeholders, and implementing the fit and predict methods for the recommendation model. They introduced features like weight and bias regularization and the integration of sampling strategies, reflecting a focus on building a functional and effective recommender system.
pythonrecommendationframeworkmachine-learningrecommendation-system
quantopian/zipline
Mar 2015 - Jun 2016
Zipline, a Pythonic Algorithmic Trading Library
Role in this project:Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:127 commits, 89 PRs, 343 pushes in 1 year 3 months
Contributions summary:James primarily focused on enhancing the Zipline algorithmic trading library with new features, most notably, a Cythonized Security class and extensions to the Asset class, creating Equity and Future classes. They refactored the codebase, renaming Security to Asset for better consistency with the addition of futures trading. Furthermore, the user made contributions to the build process by adding build_ext to the .travis.yml file.
algorithmic-trading-libraryalgorithmic-tradingpythontrading-botbacktesting-trading-strategies
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.