2023 - 2024 Distributed SQL Team

Bring TiDB, an open-source & distributed SQL database that supports transaction at scale to LinkedIn.

1. Benchmarked and tuned performance in close collaboration with PingCap, laying the groundwork to scale single TiDB clusters to millions of QPS with PB level of storage.

2. Led disaster recovery & ETL track to support the first large scale production usecase

2018 - 2023 Espresso, Core Online Data Store

Espresso is LinkedIn’s in-house, distributed NoSQL database that powers nearly all LinkedIn products at high throughput and low latency.

1. Led initiatives to support storage media agnostic backup, resulting in significant cost reductions and improved reliability.

2. Rebuilt ETL processes, achieving > 10× increase in data pipeline speed for handling petabytes of daily processing.

3. Built integrated caching mechanisms as an infra offering to reduce latency for critical user-facing features like ads, enhancing overall user experience.

• TA in Math Department

• TA for Intro to Database System in CS Department

• Grade assignments and homework

• Hold weekly office hours

• Abstract the seams between two distributed storage solutions, ESPRESSO (LinkedIn’s distributed document store) and Ambry (LinkedIn’s open source distributed blob store).

• Design and implement new client for ESPRESSO to extend its functionalities to handle large request by leveraging Ambry as its storage media.

• New client provides flexible API including streaming PUT and streaming GET in addition to standard GET and PUT.

• Interned for the Who’s Viewed My Profile team and created new notifications

• Built Android client side of new features Who View My Post / Share, created automated UI tests, and contributed to API work

• Gained deeper familiarity of the model-view-controller software architecture pattern