End to end feature implementation (index, reverse index, model ingestion)

Participated in team on-call for serving and workflow issues

Retrained GBDT and Neural Network models to add new features

Hyperparameter tuning for GBDT model

ML fairness research to reduce bias in recommendation model by skin tone (data analysis through retraining and experimentation)

Reduced time to completion of recurring compliance work from 2 months to 2 weeks via process streamlining and documentation

Conducted data investigations to find and fix causes for dips on core metrics, involving cross functional work and deep debugging across clients

Drove changes to web infrastructure and deployment process for testing standards, increasing test stability and reducing runtime while increasing frequency of testing. ( 6x tests, 10x runs a week, 10%-> 90+% reliability)

Involved in high visibility incident triage and fixing, communicating with key stakeholders and coordinating work to minimize production impact

Onboarded web engineers to test tooling at Pinterest

Led frontend framework migration, from AngularJS to Vue

Led implementation of in house component library, including initial timelines, setup, and documentation standards

Mentored junior team members, increasing their productivity and code quality (measured through pull requests/comments)

Improved development standards by introducing front end unit testing and better documentation

Core member of recommendations team: generating specs for projects, proposing solutions, and creating UI elements to showcase recommendation algorithms

Front and back end development on legacy and new projects in PHP and Javascript (jQuery, React)

Deployed applications to AWS EC2 / Heroku instances, gathered requirements, and maintained project timelines

Reduced page load time by 88% – a difference of ~9 seconds – on a legacy project through refactoring and SQL call reduction