Reduced latency on the Search homepage by optimizing JavaScript size.

Developed an internal cross-platform server-driven rendering library to update UIs without waiting for client releases.

Automated code generation for UX designers to update UIs without engineer involvement.

I worked on the File System team restructuring a feature for the administrative console, designing a new method for transferring files to avoid file corruption and locking conflicts in PHP. I created the service that allowed a user’s content to be transferred to another in an asynchronous process and by batches, which eased load on the database and generated fewer errors than the antiquated synchronous method.

I used C# to analyze public housing transaction data to pull out certain information that we were interested in. Using the data I gathered, I wrote a Node.js and React.js application to match buyers and sellers in the Diverse Solutions database - Diverse Solutions is a subsidiary of Zillow - with the buyers and sellers in the public data. This allowed us to determine whether or not Diverse Solutions closed a deal for an agent.

Second, I used Amazon Machine Learning to determine whether or not a certain user would contact an agent on the website. I gathered data from our database as well as ran certain algorithms (cosine similarity) on the data to represent it in a certain way. I generated a CSV with the data I created and analyzed and then fed that to Amazon Machine Learning. The data I made turned out to be very linearly separable, but it needs a little more refinement to move forward in the research process.

Skills Learned: Javascript (Node.js, React.js), C#