David Cochrum

Sonny's was developing its next-generation tunnel controller, which coordinates all the equipment within a car wash tunnel. Across the industry, these controllers were previously configured by a Windows Embedded edition computer terminal built into the box housing the controller. Updates to the software involved mailing a USB drive to each customer and walking them through a manual and painstaking process.

Sonny's wanted to enable configuration from any web-connected mobile device or computer and enable over-the-air updates. I was given an architecture outline and set free to build the system as I saw fit.

Since these tunnel controllers were deployed across the world with varying levels of internet connectivity, we had to build the software to be resilient, able to operate offline, and configurable via the web without requiring complex VPN setups or the like. We achieved all these goals and presented it as an industry first at the trade show the following year.

At Curio, we had issues with customers exploiting the peer-to-peer resale marketplace to buy NFTs from themselves, inflating prices and keeping the profits. I built a system that would flag resale listings matching the criteria of an attempt at fraud. The listing information would then be posted into a Slack channel for manual review by a Curio support employee with one-click buttons to approve or reject the listing.

Curio wanted to grant access to exclusive content within their Discord server based on whether a user held a particular NFT. I built a bot application using the official Discord Node.js library to respond to specific slash commands, which would direct users to a page on the Curio site where they could connect their Curio and Discord accounts.

During my time at NinjaCat, I was given ownership over a microservice built on Laravel Lumen, which responded to incoming webhook requests from Twilio and Plivo to direct them on how to handle incoming calls. The NinjaCat platform allowed users to configure dynamic call routing based on various parameters, including time of day, caller location, search engine, etc.

The response time of this API was immensely important since each second a caller spent waiting to be connected meant a significantly increased likelihood that the caller would hang up and our clients would lose out on sales leads. This service handled millions of calls monthly and needed to meet 99% uptime service level agreements.

In addition to the back-end webhook handling, I helped maintain a front-end JavaScript snippet that would be installed on client websites. This script would dynamically replace pre-configured phone numbers with a number from a collection registered via Twilio or Plivo for this client. Incoming calls to those replacement numbers were handled by our platform, which would later attempt to match to a search engine click. This gave the client a complete picture of the user journey and better sales conversion insight.