Jan 2020 – Aug 2021 · 1 yr 7 mos

• Team lead for confidential initiative.

Jan 2018 – Jan 2020 · 2 yrs

• Amazon SageMaker Notebooks and Studio
• Team lead for SageMaker Studio which launched at re:Invent 2019. Wrote the AWS blog post: https://aws.amazon.com/blogs/machine-learning/learn-how-to-select-ml-instances-on-the-fly-in-amazon-sagemaker-studio/?latest
• Built the remote kernels feature for Jupyter allowing the main Jupyter server to run on one machine with the kernels running on a remote machine.

Jan 2017 – Jan 2019 · 2 yrs

• Amazon Comprehend Document Classification
• Launched a document classification service at re:Invent. Powered by multiple algorithms including fastText, and featuring automated model selection and hyperparameter tuning.
• Submitted multiple patent applications for using model compression and AWS Lambda to load per-customer models on-demand and rapidly serve initial requests using a quantized model while waiting for the larger model to load into main memory on the service.
• Built a tool which used Bayesian optimization to optimize the number of worker threads for the webserver, the queue size between the webserver, and the number of ML worker threads.

Jan 2016 – Jan 2019 · 3 yrs

• Launched AWS SageMaker LDA algorithm and AWS Comprehend topic modeling at reInvent 2017.
• Team lead for algorithms on AWS SageMaker. High performance LDA implementation which is up to 10 times faster than Mallet.
• Went from 0 code to reInvent keynote in 9 months.
• Grew team from 1 to 6 engineers and 2 applied scientists.
• Launched AWS Comprehend topic modeling. Makes topic modeling available to all AWS customers giving them insight into their documents and data.

Jan 2013 – Jan 2017 · 4 yrs

• Launched a company wide development tool for working with NodeJS and Javascript projects. Currently used to build over 75,000 projects (version sets) and basically used on every single team in some form at Amazon and AWS. Gave regular talks and tutorials on it at internal conferences and developer workshops.

Jan 2011 – Jan 2014 · 3 yrs

• Research into higher-order correlations in the brain.

Jun 2007 – Sep 2007 · 3 mos · Woods Hole, Massachusetts, USA

• Used Matlab to analyze turbulent flows numerically simulated using Fortran.
• Co-authored paper in the Journal Of Fluid Mechanics.

Sep 2006 – Sep 2010 · 4 yrs