Quickly onboarded to become an instrumental contributor in ensuring the success of our smart glasses products, including Ray-Ban Stories. Updated and created new pipelines for building, testing and releasing new firmware that also reduced latency and increased reliability. Became the key cross-functional member on numerous projects as the most knowledgeable person to help other teams achieve their goals.

• Reduced build failure rate from 40% to 2% by identifying key breakage points and working cross-functionally to get them resolved.

• Created dashboards with alerting features to effectively communicate the current status of the pipeline and push progress towards reliability

• Decreased latency of builds from 12 hours to under 1.5 hours using metric analysis of bottle necks, increasing parallel runs, decreasing the number of flaky tests, and keeping stake holders accountable when needed

• Lead and organized offsites for the team to increase morale and help use work as a more cohesive team.

• Was instrumental in onboarding new members and convincing them to join the team helping them adjust quickly to the team and culture at Meta.

As part of the Machine Learning Experience team, I spearheaded the rearchitecture of our drive write limiting system to be simpler and easier to test in the effort ot migrate how the calculations are done. Created a simulator to determine how New Years Eve would affect our caching infrastructure to prevent outages. Also spent significant time and effort to reduce our immediate response load by implementing new tooling and lead our better engineering efforts.

• Onboarded to complex & modern C++ backend architecture and Meta tools and process

• Rearchiteching our write limiting tooling unblocked other teams to implement new caching

• Changed all our tooling to fall into a “Pit of Success” mindset making breaking production very difficult and doing the normal things easy

• Simulations for NYE used data from previous years and simulated the write/read load on our devices to see if we could handle the burst

• Migrated our ML Models over to a safe slow-deployment method to catch issues only seen in production in a safe way

• Created a series of lunch & learns to help keep the team social while working remotely and bond them as a unit.

• Put all our documentation into code to be reviewed along side the relevant diff change

I completely re-architected how Jenkins was used from the ground up utilizing Jenkins pipelines to create object oriented code that could be re-used to build, test and deploy our microservices. Heavily involved in creating a better development experience, easier workflow, and enhanced testing to make sure the product was up to proper standards.

I was the sole quality engineer responsible for the entire Panjiva platform. Including dataquality,backendtests,unittests,frontendtestsandend-to-endacceptance tests. Re-engineered the entire infrastructure to be elastic and flexible while dramatically increasing both speed and coverage.

Led a technical team to perform full stack testing with python, REST, and pytest integrating with Java via Jpype, and Java development with Junit tests to add product functionality. Oversaw Jenkins administration as it affected release engineering and test automation and solved reporting issues by creating Jira test reporting plugin for pytest. Accepted critical path for releases responsibility by taking on both the top priority features for testing and becoming the primary release engineer for product.

Key Accomplishments:

• Increased quality and efficiency of product release, reducing launch time by 90%, by immediately recognizing deficiencies and implementing process improvements.

• Improved test coverage by implementing special product features and solving multiple testing issues, including the ability to run concurrently to lack of discipline with object oriented design.

• Enhanced visibility, accountability, and metrics for tests by creating a Jira plugin that reported all results into a modular dashboard where test trends and overall quality could be easily determined.