David Pfeffer

Using Cordova and React, I did the system architecture and then constructed a trading app for simulated "stocks" representing a team winning a sporting event. Real-time market information is transmitted via websockets from the back end, and graphs/charting is done with the D3 library. An Android version of the app is now released and an iOS app is making its way through Apple review.

This app started in Angular and for a while we used Ionic as well as NativeScript. However, both turned out to not be mature enough platforms in that era to use.

Evaluated the existing platform and explained the pros and cons of reworking the existing code vs. starting from scratch, did a formal estimation of time/effort required to complete the project, and architected and designed a new platform with a new tech stack (Cordova, React, Node.js, and Express).

Managed the project, working with relevant stakeholders to ensure that goals were met. I set up and managed a CI/CD system using CircleCI and set up and maintained Azure-based infrastructure to host the platform.

Developed the entire front-end web interface (TypeScript/React) for supervisors and a hybrid mobile application (TypeScript/Cordova/React) for employees.

Oversaw the lead back-end developer and acted as secondary developer on back-end code (TypeScript/Node.js/Express).

I built a custom CMS for World Travel Group, a large multi-brand travel agency in PHP, the client's preferred language. The primary goal was the ability to integrate pricing dynamically into the site from a legacy booking engine. Also, I used HTML's "contenteditable" property to implement sophisticated WYSIWYG editing into the site. This was built in 2006, before similar editor components were available as freely available open-source libraries.

Using Ionic and then Ionic 2, I built a trading app for simulated "stocks" in the outcome of a team winning a sporting event for both iOS and Android platforms. Real-time market information is transmitted via websockets from the back end, and graphs/charting is done with the D3 library. We did an MVP release in Ionic and then ported it to Ionic 2. We were ultimately forced to move off of Ionic 2 because the platform was not released by the time we were nearing our debut, but the port was very easy to accomplish and the platform worked very well.

I built a real-time sales intelligence tool for researching information and the latest news about companies while at FunnelFire. It is intended to be used for pre-call research as well as to stay up-to-date on a company's behavior.

The engine was built in C# on a service bus architecture, with a series of ingestors running to pull data from a variety of data sources (Facebook API, Twitter API, CrunchBase, Lexis Nexis, Freebase, AngelList, web crawling, etc.) in response to messages and then to publish results as they were found back onto the bus. A company comparison algorithm then compares the company profiles found and, depending on match confidence, either merges the results into an existing company, creates a new company, or creates a potential merge for human verification via Amazon Mechanical Turk. News stories are fed through the same bus when discovered, where they land in an Elasticsearch database for future querying. A trie of all company names is stored in a proprietary self-designed database system, rapidly matching company names found in the article to help start off various push notifications and email alerts.

I'm teaching a variety of courses to undergrads and grad students at Stevens Institute, such as advanced data structures and algorithms, introductory computer science, Linux systems programming, low-level TCP/IP networking, a crash course in advanced Java for C++ developers, and software engineering topics such as QA.

I built an Ionic (AngularJS) app for iOS and Android during my time at FunnelFire. The app does not show off my ability to create beautiful apps—the app looks fairly kludgey—but rather my ability to rapidly iterate under strict time constraints. The app was built in less than a week in order to appease various sales managers who would not buy our product without there being a mobile app presence. Much of the code was shared with the front-end web application, since they're both built in AngularJS, which improved long-term maintainability.

I built the chat widget for MedChat, an Olark-like real time chat engine with medical compliance.

While at FivePM Technology, I built a real-time vehicle tracking system for shuttle buses. The vehicles carried an onboard embedded Linux platform that I developed and maintained, a cellular modem, and communicated via a special low-byte count protocol I developed to transmit vehicle position to a server I wrote that sat on Azure. We then took those positions to calculate ETA to bus stops, and responded to requests via text message and phone calls via Twilio to report the ETA to a stop or the current shuttle locations. We also delivered the location of the vehicles on a live map online, using web sockets to update the vehicle positions and interpolating between data pushes.

The entire platform was built on .NET, with Mono running on the buses and various workers written in C# and ASP.NET with SignalR powering the back end and web app.

I developed a primitive version of an "Uber" for the Hoboken, NJ taxi system. Hoboken's taxi drivers did not have smartphones and were reluctant to have a GPS device onboard due to "big brother" concerns. However, Hoboken is also only a square mile in area.

Text message-based hails indicating where a rider was located would trigger a message to be sent out to the highest rated drivers that had sent a text-based command into the system reporting that they were "online." One of those drivers could reply reporting that they are going to accept the ride. If they were the first to reply and the request hadn't timed out, they would be texted back with the location of the rider and the rider would receive a confirmation text.

We then also used Twilio to allow the driver and rider to communicate via phone or text without releasing either of their phone numbers to one another. At the end of the ride, passengers were asked to rate their driver from 1 to 5, and low-rated drivers were deprioritized in the system for receiving text hails.

The platform was built in C# with heavy utilization of Twilio. Ultimately, Uber of course took off and Hoboken gave up trying to control the internet-based hailing of their taxis.

I am in the top 2% of overall user reputation on StackOverflow.

I taught a special topics course at Webb Institute, the renowned ocean/naval engineering school on Long Island, NY. The school did not have anyone capable of teaching a complex programming course, so I was brought in to instruct the students in a crash-course in programming coupled with data structures and algorithms over the course of a semester.