Li Haoyi

Databricks

• Manage a distributed team of 6 across 5 countries in 3 timezones, onboarding 5 new engineers of varying seniority and teaching them everything necessary to be productive and effective at Databricks
• Led the Devtools prioritization process: surveying engineers on their pain points, aggregating feedback, and planning projects to improve things. Increased the average self-reported happiness score from 3.1 to 3.8/5.0
• Ran weekly tech talks with internal speakers to share knowledge, at both org-wide and team-specific levels
• Architected Databrick's distributed CI system, an elastic distributed task runner with improved stability, scalability, and usability to replace our aging Jenkins infrastructure
• Architected Databricks' elastic "Dev Box" service: on-demand, high-performance cloud development environments
• Re-architected Databricks' cloud integration testing platform from a rough prototype to a production-ready system
• Optimized and sped up many core developer workflows (compiling Scala, CI validation, etc.) by 5-10x
• Implemented the config language used throughout Databricks including its high-performance compiler and IDE plugin
• Introduced many development practices to the org: JVM profiling, structured error logging, code auto-formatting

Dropbox

• Built Python/Javascript static analysis tools, speeding work while preserving quality, preventing 50+ defects a week
• Rolled out an improved development workflow, reducing edit-refresh latency by 3x for Python and Coffeescript
• Led a team of 2-3 to destroy flaky tests; reduced the number of builds failing due to flaky tests from about 10% to 1%
• Helped manage our CI system comprising >3000 CI machines
• Mentored an intern completing a successful project (test quarantine) in his 12 weeks here
• Set up Selenium tests (> 300 CI machines) & got them adopted. Reduced bugs reaching production by 5-10 a week
• Modularized our >100,000 lines of Coffeescript, from a 'cat em all' build system to Require.js modules
• Built an auto-packager using machine learning to reduce the number of Javascript downloads by 200ish per page
• Performance improvements for most common pages, including speeding up the home page by 6x

Hands-on Scala Programming