As co-founder of GrowSF I set political strategy, run advocacy & electoral campaigns, write content, design ads, develop the website, and build our data analysis system. I wear all hats.

As Secretary of the Board at YIMBY Law, I set board meeting agendas, recorded minutes, and ensured we met the legal requirements to operate as a 501(c)3.As a board member, I approved budgets, set long-term strategy, and ensured we pursued lawsuits which furthered our goals as a housing advocacy organization.

I work on all aspects of the Google App Engine Standard runtimes, including modernization and new language support. I split my time across: maintaining and upgrading the Go environment (including the open source SDK), launching a new internal service for more easily adding new runtimes, and modernizing the Java runtime and SDK.

On this six month fellowship with Google's nonprofit arm, Google.org, I worked with a small cross-functional team to build an analysis pipeline for data from the San Francisco Planning Department. We conducted user interviews with public and private housing developers and city staff, built a data ingestion and visualization pipeline, and created a report with recommendations for how to improve their systems and the housing development process in San Francisco.

As a new hire on the Go language runtime environment for Google App Engine team, I helped add support for new versions of Go and regularly shipped new runtime versions and customer SDKs. After becoming the tech lead for the team, I led and implemented the modernization of the Go language runtime environment on App Engine Standard. I removed the legacy limitations of the runtime (which prevented lower-level system libraries from working) by upgrading our sandbox to a new, much more secure, environment and removing our patches to the standard library. I presented the keynote at GoCon Tokyo in November 2018 where I spoke about the modernization work: Video at https://crash.academy/video/533/1861, notes and a demo at https://github.com/sbuss/gocon-tokyoIn this role I regularly interacted with our customers via slack and the mailing list.

As a board member, I approve budgets, set long-term strategy, fundraise, and grow our network.

Docker and Containerization Tech Lead & Software Engineer, Platform TeamUse of Docker inside Counsyl has picked up very rapidly, and in response I've stepped up to become the tech lead for docker and containerization in general. This gives me dedicated time and focus to continue to improve our Docker-based infrastructure, including writing better tools, helping teams integrate Docker into their workflow, and investigate alternatives to our Deis cluster.We have explored Kubernetes and Mesos as more stable and feature-rich replacements for our Deis cluster, and have chosen Kubernetes. I'm now leading the buildout of a k8s cluster and the tooling around it to make migrating seamless. This includes a python service framework and best practices to make the k8s layer unobtrusive (and can thus be ignored by the majority of teams).My Dockerization efforts have led to a shift in how we develop and deploy applications inside Counsyl. Until now we haven't been following a 12-factor approach to services, and even getting a new service running used to take several months. Now teams are able to launch a new service in under an hour.

Software Engineer, Platform teamThe Platform team split from the techops team, with a shift in focus to writing software tools and maintaining a platform for the rest of our software developers. Our mandate was to make the development of software and launching of microservices enjoyable and easy.My first goal was setting up a puppet-based framework for deploying services inside Counsyl. Puppet proved to be a bad fit for this, since it's geared toward provisioning entire servers and not for easily deploying new services.As a side-project, I built out a deis cluster (a PaaS-type tool for deploying Docker containers) and started figuring out how to shift our infrastructure to a modern Docker-based system. I developed all of the initial tooling around this cluster, including deployment, making base images, a template for Django apps, and log collection and processing. This approach to service deployment quickly eclipsed the puppet approach and several teams and projects are now using it, with more coming every week.I've also written a bunch of docs and have given several internal talks on the benefits of a microservice architecture and how to do it well with Docker.

Software Engineer, Techops teamThe techops team wass in charge of building out key physical infrastructure, keeping servers online, and fighting fires! I chose this team because I only had experience writing software, but not with keeping an infrastructure running. I learned a lot about things I had never worked with while on this team.I implemented an LDAP-based heirarchical group structure with a simple API for protecting our internal views to a given team or sub-team.I took over a project (when the lead dev changed jobs) intended for large data storage with local caches, backed by S3. The files are encrypted during storage and transmission. This is being used by the lab and other teams to securely store files generated during normal operations.

Software Engineer, Billing Automation TeamI was the founding member of the insurance processing automation team. I had no prior knowledge of the insurance industry, but within three months I had built out our claim filing pipeline and automated handling the responses from the payers. On the initial launch it performed a man-month of insurance processing work in 20 minutes. We grew business an order of magnitude the year this launched, while only having to grow support staff by a few heads due to my automation efforts.I implemented a patient information auto-correction service which can automatically fix roughly 85% of bad patient insurance information without human intervention.I created optimal code set calculator which determines the best set of CPT Codes to bill insurance companies for the procedures performed. We don't need to bill for every service we provide (if we did our bills would be outrageous!), instead we target a particular price that is palatable to insurance companies and consumers, while still being profitable. Insurance companies pay different amounts for different procedures, so we automatically choose the optimal set of procedures to bill for.Towards the end of my tenure on the billing automation team we swapped in a new leader. I handed over the reigns well and the team is functioning exceedingly well.

I work on smart categories & refinements, a system to display relevant categories and product refine- ments for a keyword search. We are a Java service that gets over 300 queries per second in the US, backed by a lucene database spread across roughly 60 servers, and has an average response time of about 30ms. Our code affects every keyword search on Amazon’s websites worldwide.I am currently leading an effort to automatically create brand-specific stores for the 50,000+ brands in the Amazon catalog. This is done by detecting the existence of a brand keyword in a user’s query and restricting our search to products from that brand. It is language agnostic and extremely fast.I wrote a system to determine if a user’s search was generic or specific, in order to tailor which & how many categories to display. After exploring the data for a couple weeks I discovered a relationship between click through rates and specificity of user queries.I also internationalized our software stack. We were only running in the US when I joined. I added multiple language support and set up our stacks in Canada, the UK, Germany, France, Italy, Japan, and China.In my first year and a half I have had an intern, mentored two new hires, and started conducting phone screens.

Connectious.com was the continuation and application of my senior project from college. I created a recommender system with a mix of Python, C, and a PostgreSQL database. The system leveraged the mass of data users had constructed over their years of using the internet to recommend new websites.

I was the first employee at this small website development firm. With my help, the company grew markedly. I assisted new programmers and maintained an active programming role, as well as conducted meetings with clients on occasion. I developed in PHP and MySQL. A notable project was the development of a Google maps mashup with a local real estate company. I combined their database of property addresses with a geocoding service and Google maps to plot all of their available properties on a map.

SwampTickets was created to take advantage of a new Florida law that legalized ticket reselling. We attempted to establish ourselves as the best place to buy and sell University of Florida Gators football tickets, but unfortunately did not succeed. Reasons for failure include failing to adequately respond to the needs of our target audience (which didn't match our initial predictions), as well as a large pre-existing market that had operated quasi-legally for years before our entrance into the field.