We were a small venture backed company that set out to create an all-in-one platform for agencies which specialize in providing services for people with Intellectual and Developmental Disabilities (I/DD). Our application had note taking functionality along with the ability to schedule of both employees and clients.

Our big differentiator was authorization tracking for the services which people with disabilities use and are paid for by the state in real time. This allowed agencies to see how many hours or services their clients had used and remaining within each pay period, allowing them to optimize their business in a way that is difficult with other software on the market. Our clients also frequently commented on the quality of the frontend experience compared to competitors.

Despite getting some initial traction and users, we were unable to secure additional funding or a satisfactory acquisition offer.

I was involved in all aspects of the business, from customer acquisition and relations, product management, engineering, hiring, and more.

The tech stack used a few Google Cloud services like Storage and Google Cloud Run (GCR), Go, Typescript, React, Postgres, and Elasticsearch.

I created a web application for labeling data used by AI algorithms. The application allowed machine learning researchers and AI driven programs to create labeling tasks via API call. The tasks were put into different queues for humans that worked at AKASA to process. The different queues allowed people with different qualifications to work on different tasks.

Developed an authentication and permission service and framework for internal users and applications. Users and apps could be registered and configured from a centralized web UI

Both projects mostly used Go for the backend, Typescript + React for the frontend, and Postgres for storage. The auth framework could be used by internal services as a Kubernetes sidecar, and internal applications used Python packages where required to make using the functionality of both programs easier.

Technology I created at Terminal has been repurposed to manage a large number of virtual machines that students use while completing coursework. Embedded into lessons at Udacity are "workspaces", which can put things into the browser such as Jupyter Notebooks, an IDE, a SQL interface, 3D simulations, and more. Each workspace includes custom files depending on the lesson, and work is persisted for students between sessions. Some workspaces have GPUs attached required for machine learning. Over 500,000 virtual machines are provisioned to students each week through this system.

I designed and built the workspaces backend for provisioning, file persistence, automatic assignment grading, the secure network layer, and more. I've also built other large classroom features for Udacity, such as a spaced repetition (flashcard) feature.

Some of the main technologies being used are Go, Node.js, Java, NGINX, PostgreSQL, Elasticsearch, Docker, AWS, GCP, React, and Python.

Terminal.com hosted a service that allowed people to easily get computers in the cloud that generally included a collaborative IDE and other tools for sharing. Our original offer targeted many different kinds of users including developers and data scientists, but users in the education space such as university professors and online universities including Udacity and Codecademy used our services and requested additional features for their own needs.

Eventually we decided to refocus exclusively on educational use cases. We discovered that the technology behind online classrooms had not advanced far beyond videos and text, and we believed that more interactive learning environments could be created that would encourage people to create better online courses that students would enjoy more while learning faster. We were contracted by Udacity to customize our existing collaborative IDE tech for their online courses, which turned out to be a huge success - reportedly increasing the rate that students would complete a course by 3X!

On the back of this successful partnership, the company was acquired by Udacity.