• Technical Lead for an agile, small team focusing on a suite of critical internal product and business services

Main areas of product focus are performance and quality

Looker product contributions:

• High quality code reviews for peers, including software engineers, senior software engineers and principal software engineers

• High number of performance improvements in various parts of the product, some of which resulted in an order of magnitude or better improvement

• Significantly reduce memory overhead when writing/reading from the cache by implementing streaming in to/from the cache

• Reduce memory consumption in various parts of the product, some of which resulted in over an order of magnitude or better improvement. Significant reduction in outages as well as smoother operation reported as direct results from these changes

• Performance, architectural, clustering and threading improvements in many areas of the product

• Improved encryption and secure storage of sensitive data within the Looker product

• Increased cache hit rates for various entity types by up to 50%

• Increased cache retrieval performance for many entity types by over 2x

• Improved test suite performance and reduce build times (close to 50%)

• CDN Integration (AWS CouldFront)

• Implement SubResource Integrity (SRI) for securely loading front end assets (such as .js)

• Hundreds of bugs resolved

• High value code reviews for peers

• fixes in various SQL dialects

• fixes and updates to LookML modeling language

Software Stack: AWS, Kafka, Zookeeper, Java, Hadoop, Ruby, MySQL, Redis, Salt

Responsible for algorithm Research & Development in the Information Retrieval (IR) and Natural Language Processing (NLP) space as well as architectural design, guidance and implementation of a near-realtime stream processing system for big data.

• Improved performance (throughput) of binary classification by over 300%

• Stabilized accuracy of binary classification

• Implemented k-fold partitioning to automate testing of classification accuracy using root mean squared error (RMSE)

• Created internal tools to improve visibility for trainers on accuracy of classifiers as well as monitor operations of the existing system

• Architecting, Designing and Implementing a new system to resolve long standing issues in the current system such as maintainability, scalability, performance, failover and quality.

• Documented core algorithms and infrastructure that were previously undocumented and difficult to maintain

Software Stack: AWS, Kafka, Zookeeper, Java, Hadoop, Ruby, MySQL, Redis, Salt