Astha Gupta

Domino’s platform helps data science teams accelerate research, increase collaboration, and deploy predictive models faster. • Launched and drove adoption of Model Manager - a self serve solution for data scientists to deploy machine learning models as highly scalable, available, enterprise grade APIs. • Improved product usability, performance and reliability through a Beta program and rapid feedback and iteration. • Developed demos, field training, documentation, and worked closely with CS and Sales to drive adoption. • Developed and managed roadmap for all data science publishing features • Promoted to Lead; drove quarterly planning, handled gray product areas, developed roadmap for external communication, recruited, and mentored new PMs.

Premise is a survey and analytics platform for economists and social scientists to measure economic trends and policy impact in emerging and frontier markets. • Developed and managed roadmap for building a scalable, high quality data collection and analytics platform, particularly for econometrics. • Drove changes to mobile app UX and survey design to improve quality of customer facing data products and contributor retention - improved data quality and retention by upto 80%. • Drove new platform features and data pipelines that automate data quality checks such as detecting fraudulent images and user behavior - improved detection by 30-50%. • Worked closely with data scientists and data engineers to deliver high quality data products and interactive dashboards to enterprise customers.

Improved the end to end experience of onboarding, managing, and operating applications built on Twilio: • Launched several features to enhance account management, debugging, security and UX. • Drove redesign of the portal so it can scale to new products beyond voice and messaging. • Drove experimentation to improve user activation and reduced phone number costs by 30%.

Noom is an RRE backed startup focused on using mobile solutions for building healthy habits that help users lose weight and keep it off. I owned the food logging experience - the core habit that drives successful weight loss. Renewed focus and attention helped increase user satisfaction by 30-50%, user happiness by 20% and increased task retention by upto 30%. • Redesigned the experience using insights derived from user research -- in person observances, online tests, and in-app surveys to identify key user tasks, outcomes and satisfaction metrics • Wrote detailed specs with wireframes and worked with designers and developers to ensure right implementation • Contributed feature code and bug fixes as needed to speed up product launches and help out dev team (Android, Java, Python, SQL) • Negotiated and signed data deals with external data providers (food and nutrition data), created internal processes for creating unique datasets. Managed several independent contractors to create these datasets.

Cosmos is a ldata storage and compute platform to support analytics and processing of large datasets. I worked on optimizing runtime and memory usage of map-reduce tasks and overall cluster performance. • Analyzed important customer scenarios and cluster workloads to identify patterns for improving performance and cost optimization – improved latency by 30% for many high business value scenarios (using SQL, C#, Excel) • Defined and delivered solutions enabling Ad Center to consistently meet latency SLAs with advertisers • Prioritized and managed feature schedules and backlog, worked on advance planning with team and customers and getting sign off for proposed 6 month plans, etc. Program-managed two teams two a total of up to 10 developers

Team responsible for providing near real time monitoring and alarming of all production services at Amazon. • Tech Lead for data collection, aggregation and storage part of the pipeline. • Worked with senior engineers to design, implement and ship a better design with improved latency guarantees, availability, data quality and accuracy, scalability, and reduced operational burden. (C++, Perl) • Researched, designed and implemented several new initiatives to deal with new requirements - sampling, load balancing, tackling remaining bottlenecks and issues with interfacing systems • System is still in production