Back-end Developer in Thornton, CO, United States

A Go-based application where freight loads are automatically scheduled at both pickup and dropoff based on different pieces of data such as facility hours and scheduling methods. The architecture for this project was event-based due to having to pull from multiple data sources and needing the ability to retry without having to make the same service and API calls repeatedly. Kafka was used to achieve the retry capabilities and pass off information at any given step, making the architecture event-based. Also, MySQL was used to store the history of freight loads passing through the routing engine.

The routing engine was a massive success, which led to over a 50% increase in automated freight loads, and is currently scheduling 4,000 loads a day. The percentage keeps growing, and soon a majority of freight loads will be running through the routing engine at Uber Freight.

A Java Spring-based project that would ensure two different data sources are entirely in sync. The two different data sources for this specific project to keep in sync were a MySQL database and an IBM mainframe. Some teams at USAA only had the capability to read data from the mainframe, but the mainframe was no longer the source of truth, so a new sync process had to be made to ensure data quality.

This new process would attempt to update both MySQL and the mainframe, but unfortunately, the mainframe would sometimes fail in updating without throwing an error. Once the updates were attempted, a Kafka event would be emitted so that a consumer would pick up that message and try to compare the data between the two data sources. If the two data sources were out of sync, then another update would be made to the mainframe, and the same Kafka process would be kicked off again. If the same piece of data failed to sync three times, then a Slack message would be sent to our team, so troubleshooting could begin on why the data could not be synced to the mainframe.

A Go-based infrastructure that would be used by the entire engineering department at American Express. I worked on two major parts within this infrastructure, the first being the streaming of purchases made by cardholders of American Express and the second being able to call various APIs to make the correct decisions on having points added to a given customer's account.

Kafka was used for moving data throughout the new infrastructure to stream the events. An internal custom Kafka library was made in Go, which could handle all the rich features available at the time, such as retries and custom headers.

Various APIs had to be hit to apply points correctly to customers' accounts to make correct decisions. Some of these APIs contained large amounts of data, so the data had to be cached if there was a failure. Of course, due to using the custom-made Kafka library, retries were automatically built into the rest of the infrastructure.

Made a Python 3 application to scrape various airline services in an attempt to find the cheapest flights for destinations I wanted to travel to. The script was deployed to Heroku and would run every five minutes to scrape various sites. It would use various NLP libraries to gather important information such as price, dates, travel to, travel from, and layovers.