TopOff has become the premier application in GSF, and a rebuild is on the way. I strategically collaborated with a team of 30+ to build TopOff 2.0. TopOff 2.0 will be built in CDK and maintain all of the original features. It will also be heavily modularized to help with optimization and continue to serve our ever growing businesses. This includes machine learning to help with more accurate shift selection for individual stores, especially stores with low headcount. TopOff 2.0 is one of the first Tier 1 applications to come out of GSF, and will continue to staff over a billion dollars annually throughout 900+ stores.

In addition to my TopOff contributions, I built the SmartParking application that allows Amazon to remotely monitor parking lot capacity using Bosch and AWS services. SmartParking utilizes Kinesis video streaming, AWS Rekognition, DynamoDB, and Simple Notification Service to deduce the capacity of a parking lot and notify the appropriate individuals. Implementing this application significantly reduces store managers’ time spent checking their parking lot. In order to architect this solution, I had to compete against the Bluetooth Beacon team and Rabbit Gate Breaking team in a bid to build. After carefully planning and winning the bid, I was able to work closely with Bosch and AWS internal teams to execute my vision and enhance existing systems.

Further improved Deadbolt, a hardware-software stack that helps track device usage and returns throughout a physical store. This involved writing directly on hardware in Java and building a website in React. Interactions with the locker are done through MQTT and IoT, which also aligns with my LED System personal project. The frontend utilizes AppSync GraphQL to interact with the backend, creating a complex NoSQL DynamoDB database. Our team of 12 were responsible for building out the hardware, writing the underlying software, and installing the lockers.

As a BIE IC I build the Fullstack web application, TopOff, the premier staffing program for Amazon Prime. TopOff intakes large scale general data, runs complex algorithms to make the data more easily digestible for the end user, and displays a visual representation of staffing in graph form. The complex data is able to be updated by an end user and I deconstruct the data, update the backend, and deliver an updated visual within a second. TopOff has no concept of the data it is handling, since the 10 businesses (800+ buildings) it covers are all different. Instead I use detailed configs to understand what algorithms need to be run and how to handle the outputs. I also built a shift solving service that staffs 90% of all Amazon Prime Employees throughout our various businesses. By taking the number of orders and the average pace of work I am able to determine the quantity of coverage needed throughout every minute of a day. To turn coverage into workable shifts, I wrote a shift selection service, which ranks the quality of shift and selects the optimal options. Lastly, I built the automation service, which allows for solutions to automatically be generated as many times as the end user wants and allows them to set how much to cover through a percentage. This creates the opportunity to trickle staff to select employees and make sure that 100% of the needs are covered by day-of. Managers are notified of the solutions by either email or Tableau and they have the option to opt-out of the solutions and run them manually. Turning an active system into managed by exception.

To build this behemoth I worked closely with data analysts to ensure accuracy in algorithms. Labor Planning to help understand the core problem and how to orchestrate a scalable solution. Project and store managers across all businesses to understand their needs. As well as, Database Engineers to help build materialized tables to keep the application running quickly.

Updated curriculum to fit industry standards. Held 4 hours lectures on Sundays with over 30 students maintaining an average 98% retention rate. 120 course reviews with a 96% approval rating. Weekly interview prep and office hours were attended by at least 50% of the students. Created video recordings for students to effectively prepare for their upcoming interviews.

My course contained 3 parts; HTML5, JS6, and flexbox. It focused on building a boilerplate to help students attack interview questions. It's common for blank pages to cause anxiety during interviews, so providing a constant starting block can be beneficial. I will say more importantly I taught the students how to be personable during an interview. The students were already developers attempting to move into FAANG companies. Teaching them how to highlight their skills and be engaging was as impactful as the lessons, based on reviews.

Revolutionizing the wine industry.