History
As of September 2024, I'm a Software Engineer at Anthropic! It's really cool to be part of a tech industry that's growing so quickly, and I get to work with really smart people who genuinely care about the safety and ethics of AI. I work on an internal tools team to build things that allow us to scale our productivity.
Before that, I was a Senior Software Engineer at Iterative Health who puts technology to use for the benefit of gastroenterology. Detecting colon cancer earlier through computer vision and employing AI to speed up and improve opportunities for patients.
And before that, I was a Software Engineer at Amazon from 2014-2022 in:
- Order Fulfillment backend systems — we made sure you got your order when we promised by optimizing where the items come from, how they're transported, and so on.
- Ultra Fast Fulfillment Warehouse Management Systems — improving and expanding the UFF (Prime Now, Amazon Go, Amazon Fresh) software to support new verticals.
Accomplishments
Some things I'm proud of:
- Being a Built By Girls mentor and helping bringing the gap that young women face in tech
- Developing many developer productivity tools behind the scenes at Amazon such as a graphical interface for comparing A/B test results in CI, a daemon that integrates with a browser extension that adds quick links on internal sites to quickly launch shells to fleets, a script that build a full IntelliJ project from a service, which pulls the autobuilt packages with the option of also pulling in the source for other related dependencies and marks all the files correctly to allow IntelliJ to understand the project structure ‘out of the box’
- Implementing features in Amazon warehouse management systems
- work assignment tweak that allowed better
- a distance-based pick planning heuristic that reduced the amount of walking warehouses workers needed to do by 1 mile per day
- making a breakthrough in the investigation of a memory management issue in fulfillment planning that allowed us to reduce our expected fleet scaling by ~30% before peak season
- developing a user-friendly deployment script that connects multiple data sources such as ECS, git, kube to show a preview of what is already deployed, what is going to be deployed, version differences, and links to wikis that should be updated after deployment. it builds things in parallel, has progress bars, names things according to our internal conventions, checks all prerequisites before proceeding and gives warnings if anything looks out of the ordinary and all around just makes it hard to do the wrong thing.
- lead the charge on using llms to pull out and present key information about patients from their health records that allow specialists to give more time to screen patients for clinical drug trial eligibility, raising the number of potential patients we can help connect to effective treatment