Pat Ferrel
Chief Consultant at ActionML
Seattle, Washington, United States
Summary
🤩
RockstarPat Ferrel is a seasoned machine learning engineer and chief consultant with 13 years of experience building production-scale recommenders and big-data ML systems from algorithm design to containerized deployment. Based in Seattle, he has deep hands-on expertise in PredictionIO, Apache Mahout and Spark—contributing to Mahout and refactoring PredictionIO into microservices—and authored the open-source Universal Recommender implementing a Correlated Cross-Occurrence algorithm. He blends research-quality algorithm work (MMR/MR refinements, large-scale correlation methods) with practical infrastructure skills in Scala, Java, Python, Docker and orchestration. Known for helping organizations turn clickstream and content signals into actionable personalization at scale, he often bridges the gap between messy data sources and production-ready ML services. An early startup CTO and long-time committer, Pat pairs pragmatic engineering leadership with a proven track record of shipping complex recommender systems.
13 years of coding experience
14 years of employment as a software developer
Stackoverflow
5,703reputation
325kreached
133answers
69questions
javascript
top-5%
angularjs
top-5%
apache-spark
top-5%
Github Skills (33)
algorithm10
hbase10
algorithms10
mahout10
predictionio10
back-end-development10
amazon-elasticsearch10
big-data10
build-system10
data-structure10
java10
scala10
javas10
elasticsearchquery10
aws-elasticsearch10
Programming languages (9)
Github contributions (5)
actionml/universal-recommender
May 2015 - Jun 2019
Highly configurable recommender based on PredictionIO and Mahout's Correlated Cross-Occurrence algorithm
Role in this project:
Back-end Developer
Back-end DeveloperContributions:249 commits, 21 PRs, 69 pushes in 4 years 1 month
Contributions summary:Pat contributed to the implementation of core backend features for a recommender system built on PredictionIO and Mahout. Their commits show the initial development of the DataSource, Preparator, and Engine components, suggesting they were involved in setting up the data ingestion, preparation, and overall system architecture. Additionally, the user refactored code related to the MR (Maximum Relevance) to MMR (Maximal Marginal Relevance) algorithm and ES-based indexing, indicating responsibility for a core part of the system. The user was likely working on the core logic and integration of external systems, with some initial data model work.
mahoutconfigurablerecommendercollaborative-filteringpersonalization
apache/mahout
Jun 2014 - Jun 2018
Mirror of Apache Mahout
Role in this project:Back-end Developer
Back-end DeveloperContributions:50 commits, 21 PRs, 104 comments in 4 years 1 month
Contributions summary:Pat's commits primarily involve modifications within the Apache Mahout project, focusing on enhancements and refactoring of existing code. These changes include updates to the Conjugate Gradient Solver, distributed algorithms, and various Java-based components. The user demonstrates a good understanding of the project's mathematical foundations, data processing capabilities, and Hadoop ecosystem.
javamahoutapacheapache-mahout
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.