Senior Data Scientist
2019 - 2021CVS Health- Served as lead data scientist on various machine learning development, experimentation, and workforce innovation projects providing the incremental annual value of XX million USD.
- Responded to consistent urgent requests from the White House, CDC, and Operation Warp Speed leadership on capacity planning, second-dose adherence, and daily vaccine utilization.
- Implemented a fully customizable suite of COVID-19 vaccine demand forecasting and demand diagnosis models. These models anticipated vaccine demand drops and highlighted potential areas for intervention.
- Collaborated with lead designers to identify and address the impact of social determinants of health on low immunization rates, drawing on SHAP values and ethnographic data to design interventions.
- Acted as a key contributor to CVS's enterprise-wide algorithmic bias policy which set out steps for monitoring and mitigating bias along protected class lines within AI systems.
Technologies: Artificial Intelligence (AI)Data Scientist
2018 - 2019Widget Brain- Led retail projects including demographic-based demand forecasting for a large supermarket, roster optimization for a large Australian cosmetics chain, and theatre attendance forecasting for a large Australian cinema company.
- Delivered predictive maintenance models for a large shipping OEM, allowing a 66% reduction in sensors.
- Implemented deep learning extensions (such as LSTMs) to the existing demand forecasting product.
- Built production flows using NodeRed and deployed models using AWS serverless code tools.
Technologies: OR-Tools, Node.js, Jupyter, PyTorch, PythonResearch Officer
2016 - 2018Australian National University- Built NLP machine learning models to predict the likely severity of identity theft case reports. Research officer on Australia's first large-scale study on identity theft.
Technologies: Jupyter Notebook, PyTorch
Machine Learning Developer in Boston, MA, United States
Member since June 17, 2019
Michael is a data scientist and machine learning engineer. Most recently, he led CVS’s COVID-19 vaccine demand forecasting, liaising closely with the White House and the CDC as part of Operation Warp Speed. He spends his spare time working on AI ethics problems and is an advocate and mentor for queer AI developers. An experienced leader, Michael has overseen teams of data scientists on health, workforce, and industrial projects.
Michael is now available for hire
Portfolio
- CVS HealthArtificial Intelligence (AI)
- Widget BrainOR-Tools, Node.js, Jupyter, PyTorch, Python
- Australian National UniversityJupyter Notebook, PyTorch
Experience
- Data Science 8 years
- Python 8 years
- Machine Learning 8 years
- Natural Language Processing (NLP) 8 years
- Pandas 8 years
- PyTorch 5 years
- PySpark 3 years
- Spark 3 years
Availability
Part-time
Preferred Environment
Amazon Web Services (AWS), AWS, Git, Python, Natural Language Processing (NLP), Spark, PySpark, Computer Vision, Machine Learning, PyTorch
The most amazing...
...project I've coded is a demand diagnosis models to understand reasons for COVID-19 vaccine hesitancy across the USA. It saved hundreds of lives.
Employment
Experience
- Generalized Demand Forecasting Model
Together with a team of data scientists, implemented and used a data forecasting suite including over 20 different models
- Supermarket Demand Driver Model
Lead developer of a machine learning model using census demographic data to predict the success of supermarket promotions, expansions, and luxury items in a given area.
- Operation Warp Speed Demand Forecasting
• Implemented a fully customizable suite of COVID-19 vaccine demand forecasting and demand diagnosis models. These models anticipated vaccine demand drops and highlighted potential areas for intervention.
• Responded to consistent urgent requests from the White House, CDC, and Operation Warp Speed leadership on capacity planning, second-dose adherence, and daily vaccine utilization.
• Collaborated with lead designers to identify and address the impact of social determinants of health on low immunization rates, drawing on SHAP values and ethnographic data to design interventions. - Machines and Trust: How to Mitigate AI Bias (Publication)Unwanted AI bias is already a widespread problem. Machine learning models can replicate or exacerbate existing biases, often in ways that are not detected until release. So what can be done about it?
Skills
Languages
Python 3, SQL, PythonFrameworks
StrongLoop, SparkLibraries/APIs
PyTorch, Pandas, PySpark, TensorFlow, Facebook API, Keras, Scikit-learnParadigms
Data Science, Agile Software DevelopmentPlatforms
Jupyter Notebook, Amazon Web Services (AWS), AzureOther
Convolutional Neural Networks, Machine Learning, Artificial Intelligence (AI), Computer Vision, Natural Language Processing (NLP), Neural Networks, Deep Neural Networks, LSTM Networks, AWS, OR-Tools, GeoPandasTools
Git, JupyterStorage
MySQL
Education
- Graduate Diploma in Computing2016 - 2018Australian National University - Canberra, Australia
- Bachelor's degree in Law2013 - 2016Australian National Unviersity - Canberra, Australia
