Solving scheduling, workflow orchestration and big data abstraction problems at scale as part of data platform engineering.

Tech Lead, Ads Funnel Performance tradeoffs

• Helping make experimentation analysis around latency/system metrics easier through better integration with ads A/B testing framework

• Identifying opportunities to improve funnel performance latency/save infrastructure costs

Tech Lead, Data Sharding Solution

• Lead the design, planning and development efforts of a generic horizontal data sharding solution for

services at Twitter working with multiple cross functional teams across various pillars. Blog link - https://blog.twitter.com/engineering/en_us/topics/infrastructure/2021/sharding-simplification-and-twitters-ads-serving-platform

• Worked with Core System Libraries to open-source some zookeeper serverset announcement logic and

build foundational support into platform as part of the sharding effort.

Ads Databus Team

• Designed and developed a solution to scale the delivery of Ads MySQL DB updates to read-side

applications with a lambda based architecture using Kafka for streaming and HDFS for snapshots.

Ads Ranking Service Team

• Founding engineer of the Ads Ranking Service team involved in design, prototyping and launch of a

new micro-service split from the monolith Adserver for running late stage ads ranking funnel. Blog link - https://blog.twitter.com/engineering/en_us/topics/infrastructure/2020/building-twitters-ad-platform-architecture-for-the-future

• Performed various capacity planning and Java/Scala/JVM system optimizations to reduce

infrastructure and hardware cost of launch for Ads Ranking service.

• Designed and Implemented a distributed cache layer for decayed ad engagement metrics moving from the older in-memory caching solution.

Cross DC Spend replication

Blog link - https://blog.twitter.com/engineering/en_us/topics/infrastructure/2020/how_we_fortified_twitters_real_time_ad_spend_architecture

Graduate Teaching Assistant for "Operating Systems", "NET Web application development" and "Special Topics - Applied Internet Technology".

AAA - Account Activity API Engineering Team

Technologies Used - Scala, Ruby, Apache Mesos, Apache Aurora, EventBus, Finagle, Twitter Server, Finatra, Thrift, Wiley, Apache Zookeeper and Manhattan.

AAA Delivers Twitter data like Tweets, DM's, Follows, Likes to customers who have subscriptions with us via Webhooks via a real time API to enable a rich developer ecosystem, for example, building AI chatbots.

• Improved end to end latency for delivering like events by reducing the number of EventBus hops in the AAA pipeline.

• Owned the design and implementation of scaling the AAA service before going into production

• Designed the delivery of new events - Blocks and Mutes via the AAA service.

• Moved the storage system from Manhattan(distributed key-value storage system) to MYSQL for storing webhook configurations and cached them in Manhattan for faster read performance.