Member of the Kubernetes Autoscaling team at Databricks, working on systems that improve efficiency, reliability, and scalability of large-scale data workloads. Primary focus is Autopilot, an intelligent autoscaler that leverages historical workload patterns to dynamically adjust CPU and memory resources.

Key areas of contribution include:

Efficiency optimization through intelligent VM instance selection and management

Cluster health monitoring to ensure stability, scalability, and cost-effectiveness at massive scale

Enhancements that enable Databricks customers to run data and AI workloads seamlessly and efficiently

Dedicated to building cloud-native infrastructure that balances performance, reliability, and cost savings for enterprise-scale data platforms.

• Implemented log parser in Python to parse trade logs into KDB database, using Q to query the database and calculate stats such as fill % and return results to the user interface

• Built an analytic dashboard in React to show linkage through trade process from client to market, allowing traders to easily query client orders and discover where slippage was occurring

• Collaborated closely with traders to gather user feedback and requirements, translating their input into a user interface

• Ported over a rules service engine to follow new language specifications, allowing new detections to be run for a user behavior threat detection software

• Used Java to write and JUnit test new compiler functionality, with over 92% code coverage

• Added support for JSON structures and implemented optimizations to flatten AST involving nested field extractions

• Implemented message aggregator in Java to reduce the number of shipment capacity breaches, using JUnit and Mockito to perform unit testing

• Used Kinesis to manage parallel data streams to avoid hot partition problems when updating DyanmoDB with updated shipment information

• Implemented thread safe code to allow several hosts to run in parallel, processing over 2000 shipment updates a second

Implemented interface and functionality using TypeScript with React for user account and influencer pages for an e-commerce website

Developed scripts using GraphQL to query and mutate data as well as storing and updating the cache directly, leading to a 40% reduction in network traffic

Implemented Google Analytics tracking throughout the site to track user interaction as well as API calls