Professional Biography
My mom showed me my first lines of Basic when I was 5 years old and I have been fascinated with software ever since. It was a simple for loop that used *’s to output half a Christmas tree. It looked something like this.
| |
I started writing my own code when I turned 11 by putting together simple HTML pages to showcase my favorite cartoons and video games. Soon after learning HTML I picked up Perl to make my websites dynamic, and then Python after that, using both to build websites with CGI scripts.
At 14 my mom was able to get me a job in data entry for the Survey Research Institute at Cornell. I automated the collection of addresses and phone numbers using Python scripts to scrape whitepages.com. This got me a job working with the web development team for the School of Industrial and Labor Relations, where I was able to branch out into learning ColdFusion and “dynamic HTML” with YUI and CSS.
My first entrepreneurial venture was to write the website for a local theatre company with my friend Kevin Cheng. We put together a decent website and got paid $300 to do it. It was great fun.
In college, I studied biology with the intent to go into medicine. After 3 years I learned I probably did not actually want to become a doctor, so I took several CS classes in addition to the biology classes. After I graduated college, I was able to live abroad in Qatar. In Qatar I was a teaching assistant for organic chemistry and biochemistry at the Weill Cornell Medical College in Qatar.
Once I got back to the states, I moved to San Francisco to work with my sister at a software consultancy she helped to start called Blazing Cloud. One of my first contracts was with a company called TrueCar, where I was indoctrinated into Bay Area tech culture and development practices from Pivotal Labs. TrueCar is also where I brought my first ML model to production with Mikhail Semeniuk. It was a use-car pricing model, prototyped in SASS, and brought to production first in JavaScript, then in Java when it needed to run quickly.
After Blazing Cloud, I worked at Linden Lab in the new products division. I was able to work on Versu, which was a virtual storytelling game where the goal was to interact with other characters powered by AI. I was also able to work on Blocksworld, which was a building game where you could create and program almost anything you could think of.
After Linden Lab I went back to TrueCar, this time as a full-time employee. Again I found myself working on building pricing models with Mikhail. We go the idea to build MLeap after deciding it was too difficult a process developing ML models then bringing them to production as API services to power products.
After 2 years at TrueCar, I left to start Combust.ML, which was a company built around the commercialization of MLeap. I co-founded Combust with Mikhail. Combust was able to pay the bills for a while, but it became apparent pretty quickly it would not be a wild success. So, Mikhail and I co-founded another company, Propel AI, along with Tom Taira and Bernie Brenner.
We spent 1 and 1/2 years building a conversational AI at Propel to sell to dealerships. After that time, we sold Propel AI to Carvana to integrate with their systems. When we went to Carvana, there were 5 of us. When I left Carvana, we had grown the team to over 30 people including software engineers, product managers, data scientists, and data labelers. Propel AI software was (and is still) powering tens of thousands of conversations a day, connecting Carvana customers to the AI we built and the customer advocates when the AI isn’t enough. After 3 and 1/2 years at Carvana, I decided to take some time off to explore opportunities outside of the automotive industry and hone my skills by writing open-source software and technical articles.
It took 1 month before I picked up another job working in the BioTech space at CytoTronics. My role was Principal Cloud Engineer, which meant I was setting up AWS and Kubernetes infrastructure, writing the API for data management, integrating our device orchestrators with the cloud, and building notebooks for our biology team to analyze the data. My role brought me into the space of data analysis to help interpret the data we were collecting. I supported the biologists in running their analyses, and I was able to run my own analysis of distributions of cell cultures in wells over time. The results were published in a paper and can be found in Figure 4: A semiconductor 96-microplate platform for electrical-imaging based high-throughput phenotypic screening.
I left CytoTronics towards the end of 2023 to go back into the world of consulting and pursue a lifelong dream of mine to learn how to build video games. I am currently focused on learning graphics rendering and game development using Rust and the open-source Bevy game engine.