- Collaborated with another team and cross-functional roles and implemented APIs and various backend components for a redesigned merchant catalog ingestion workflow aiming to improve operators’ efficiency and reduce human errors from working with spreadsheets. Received positive feedback after a successful demo of the prototype.- Extended photo formats officially supported by DoorDash to increase operators’ efficiency, reducing the time needed by 90% for 40% of merchant-provided photos’ ingestion.- Improved photo uploading workflow used internally at DoorDash to catch broken photo links at ingestion time, preventing more than 55K broken photo links from being ingested.- Refined our service’s dashboard and metrics collection from on-call experience, improving on-call efficiency and making issues more easily noticed.

- Designed and built a data loader system to ingest weather data from multiple sources with different file formats and store them in a data lake for downstream use cases, and collaborated with scientists on building ML models that respond to bad weather forecasts and protect customer promises by padding promised delivery dates. Integrated our model with Amazon Production APIs to adjust customer delivery promises in forecasted bad weather areas, saving the company ~5 million USD per year.- Designed and built a data pipeline to publish pre-processed weather data along with our model outputs to Amazon's internal data platform so that other teams can also have easy access to formatted weather data for their use cases. Worked on a design for a scalable service that vends weather data and varies model outputs on demand through APIs to allow engineering teams to consume data through APIs instead of from data warehouse tables.- Designed and built a website that visualizes customer delivery protection configurations with interactive maps and filters. Iterated on both front-end and back-end implementations to add new features requested by customers throughout peak season while also improving overall website performance by ~5x compared to early versions.- Generalized the cloud infrastructure used for the website into a demo project for Amazon internal engineers to build up their websites quickly and easily. Documented all of our lessons learned in 4 months into an internal Wiki page, and got more than 1000 views within the first 3 months. Made further improvements on it and presented it at internal WebDevCon as a workshop session and got positive feedback.- Built an Android Java application on an IOT camera device for tracking dwelling pallets in Amazon warehouses in collaboration with the Amazon Robotics team. Designed and implemented backend service APIs using AWS Lambda to send data from/to the cameras and store them in S3 and DynamoDB for our web portal to display.

- Worked with customers in Amazon Fulfillment Center to build a dashboard that helps catch delays in each step of an order’s lifecycle to avoid delivery estimate misses. - Redesigned existing SQL queries to pull data from live data sources instead of data warehouses, enabling near real-time reporting. Improved query runtime by ~10x to make it feasible for real-time analysis.

Worked with other group members to design new online courses for users.