Owned and implemented multiple fullstack projects that involved taking extracting, analyzing and transforming patient data so that we could find good candidates for clinical research trials.

A typical project would include:

One major project was a complete rewrite of a video redaction service; it took video data and using AWS Rekognition it would detect and censor any personal information. This system was rewritten to do this work in parallel and was 3x as efficient afterwards.

Development of new features and internal architecture for products used by large enterprise customers. These features have to process TBs of data per day for thousands of customers without breaking or any performance degradation.

To ensure that we hit this level of quality, I owned the projects completely. This includes writing documentation, leading kick-off meetings, writing test plans and test suites, and monitoring performance after deployment. In some cases I was able to guide junior engineers how to use industry best practices and improve their testing and debugging skills.

Example features:

API architecture and frontend design to export profile and company data to marketing sites such as Salesforce and Hubspot.

Rewriting bulk pipeline logic to achieve a 6x speedup while lowering resource usage.

Adding SingleSignOn support via SAML.

Worked on large scale data processing for Tenor at Google.

Before I left I had created a pipeline that can take new uploads from user generated content and:

1. Transcode the gif into a variety of formats

2. Upload these new formats to our legacy DB and the new Google backed DB.

3. Use Machine Learning to annotate the images with tags to mark them as NSFW, featuring celebrities, animated, etc.

4. Handle logging, monitoring, alerting and other related systems while this happens.

Software engineering with a focus on internal backend architecture. Major features include:

Tenor Insights, a tool for curators to manage top trending gifs.

Neon is a company that specializes in image processing, mainly in the realm of thumbnail generation.

From a technical perspective, I worked on the internal tools used to process all the requests. This includes:

Tech Stack: Java, Python, AWS, Chef, Tornado, HBase, flume, REST

From a non-technical perspective I was able to: