Saratoga, California, United States
I build and lead teams that tackle complex, large-scale challenges in everything from AI, to distributed systems, to programming languages. Transforming poorly-defined problems into finished solutions really motivates me. I’m comfortable working across different levels of abstraction: implementation-level, team-level, organization-level. I like working with teams that seek out data and second opinions, to refine and improve ideas throughout the development lifecycle.
I build custom AI models that identify short-term trading opportunities and portfolio optimizers to exploit those opportunities. I use state-of-the art AI techniques to transform raw technical, fundamental, and econometric data into predictions. I’m focused on equities trading.
My team makes the common infrastructure Google developers build services with. Our customers make products you've heard of, like YouTube, Search, and Maps. We focus on things like reducing development toil and making services more reliable, safe, performant, and scalable. We’re responsible for the frameworks for building services. Our current focus is modernizing the stack to account for rapid changes to the development life cycle, powered by the rise of generative AI.
Lead a 15 engineer team supporting the Python programming language at Google. We supported our AI customers (DeepMind, TensorFlow, etc.), as well as other large users of Python like YouTube. Our biggest accomplishment in my tenure was migrating the world's largest Python codebase from Python 2 to Python 3. This was a huge effort involving cooperation from hundreds of teams across Google. Through automation, my team was able to save Google hundreds of engineer-years worth of toil as we migrated the code base.
* Led Wealthfront’s data and machine-learning engineering team, growing the team from a single engineer (me) to twenty people * Led buildout of Hadoop, and later Spark, ecosystem, including ETL jobs, data warehouse, dashboards, near-real time system, etc. * Built a reenforcement learning system (via Bayesian inference) to launch and ramp product experiments
I worked on several infrastructure projects at LinkedIn. The one I'm most proud of was called Quick Deploy, a project which changed the way we wrote code at LinkedIn. It lead to a 20% productivity boost across the engineering organization at LinkedIn. I also worked on frontend infrastructure at LinkedIn, and I briefly managed LinkedIn's i18n team.
While at Agilent, I was the lead engineer behind Command Expert: http://tinyurl.com/AgilentCommandExpert This is a PC application to integrate electronic test integration into Excel, Matlab, LabView, and other analysis applications. I came up with the basic idea, worked with customers to make it better, and implemented most of the UI. Prior to that, I worked on programming languages and user interfaces for distributed, real-time systems. This work was focused on electronic test customers who were using IEEE 1588 for clock synchronization in measurement applications: http://www.nist.gov/el/isd/ieee/ieee1588.cfm