• Collaborated with scientists, engineers, and RAs to better structure, manage, and analyze data

• Developed 30+ Python applications on rapid timelines to cover last-mile data management needs, prototypes/proof-of-concepts, and one-off software solutions, which reduced time spent managing data by up to 4 hours per user per week for each tool, while also enabling previously impossible workflows, structuring data more efficiently, and reducing manual errors.

• Worked directly with end users throughout the product development lifecycle - gathering requirements, agreeing on a product design, incorporating feedback, implementing bug fixes and feature requests.

• Architected system for deploying user tools as Jupyter notebooks that streamlined both user and developer experiences (and reported on details in a blog post here: https://www.zymergen.com/blog/technology/building-a-platform-for-duct-tape/ )

• Created course materials and led training courses to educate scientists on applying Python, MySQL, and Github to their work. Around 50-100 employees took the courses, which allowed them to better investigate their own data and write their own processing scripts for simple tasks.

• Developed and organized various team-wide processes: authored team charter; organized several team-wide code cleanup and documentation initiatives; developed framework for levels of software implementation of user requests.

• Acted as liaison between end users and software teams: escalated potential technology gaps to software, handed off short-term solutions that needed to grow into full-production products.

• As a senior member of the Ad Delivery team, supported and improved the backend infrastructure for displaying ads to large scale live traffic

• Discovered and corrected many root causes of revenue loss due to concurrency errors in a distributed backend stack; reduced error rate from thousands per day to tens per day

• Constructed pipeline to aggregate budget fulfillment data from various sources into cloud-based database; automated data joining process for more efficient monitoring and investigation

• Participated in team-wide mutual code review for all code changes; pushed them to production with zero downtime on machines serving live traffic; monitored and triaged these services 24/7 as a member of on-call rotation