- developed in-house clinician scheduling tool with 1 peer- led AWS-backed infrastructure design and implementation using aws-cdk, s3, RDS, lambda, apigw- designed and maintained CI/CD pipeline for scheduling tool using aws-cdk, jenkins, docker- ported scheduling algorithms previously running in sagemaker to a lambda api backend- developed frontend features on scheduling tool in React- mentored summer intern- supported devs across teams working with s3 bucket deployment and docker-backed lambdas

- Designed and implemented serverless ETL flow using AWS CDK- Designed and implemented dataset and model versioning system backed by S3- Developed and dockerized inference python code supporting Keras and Caffe models- Trained and deployed image classifiers in Keras for use on surgical video- Mentored junior coworkers in software development and ML project management- Managed the data science sprint planning, backlog, incoming request triage

- Led ground-up re architecture of the data science code base from loose scripts into modularized, reusable, unit tested code with coding standards (Python, Git, pylint)- Ported fully custom NLP service from contractor env to in-house- Developed system for crowd-sourcing (MTurk) annotation of video- Rebuilt image training data generation system to be 100% faster by leveraging AWS services- Primary developer for offboarding and onboarding of contracted ML dev teams

- Developed user-facing features as a full-stack software developer (JavaScript, Mongo, Meteor, AWS)- Acted as primary developer contact between C-SATS and ML contractors- Designed and implemented training data annotation and preparation pipeline- Deployed several open source annotation tools (CVAT, Vannot) as well as custom internal tooling

Sanoma is a leading media group in the Nordic countries, making it an ideal place to study large amounts of data. During my time there, I developed and deployed a full data pipeline and recommendation systems for my masters thesis on serendipitous recommenders and the filter bubble effect.• Implemented data pipeline and recommendations algorithms on a live platform working with AWS, Hadoop, Jenkins, Apache Spark, and PredictionIO machine learning server• Used ELB and Monit to create redundancy and increase robustness to achieve 100% uptime for month-long experiment serving over 45,000 recommendations daily

The Instructional Technology Services department at UCSB is responsible for the Course Management System (CMS) used by faculty and students. Beyond supporting typical use cases of a CMS, our software team explored learning analytics. We built applications to surface information on how the CMS was being used to instructors and students to improve their learning experience.• Developed several Javascript applications with the ExtJS framework, involving responsive design, JSON data, MySQL backend, and PHP scripting• Worked with supervisor to design intuitive ways of aggregating and displaying data

Invoca is a Voice Marketing Cloud platform. The platform brings phone calls into the digital age, making calls as measurable and automated as clicks.• Worked with team of Ruby on Rails developers using TDD and scrum methodologies• Participated in daily standups and development sprints including feature development, testing, and quality assurance• Gained experience with RightScale cloud deployments and operations