Rinse is a technology-enabled dry cleaning and laundry delivery company. We built a polished website and cross-platform mobile apps, servicing tens of thousands of customers in five metropolitan areas. This includes A/B testing, a CMS, technical SEO, years of iteration on signup funnels, and intensive instrumentation via Segment and Amplitude. We built this on a Django backend, with a modern Webpack + Sass + BEM frontend.
We built Rinse as an SMS-first business. Our primary customer interaction is over SMS - they can schedule pickups and deliveries, interact with customer service, get status updates, and be re-engaged over text message. We built marketing automation tools, customer service integrations, and automated response mechanisms, sending and receiving tens of thousands of SMSs per day. This was built on Django with integrations with Kustomer and Twilio.
The majority of the technical complexity at Rinse is in fulfillment, with tools for delivery drivers, warehouse personnel, dry cleaning and laundry vendors, dispatchers, quality control personnel, customer service, and management. Some are mobile apps, others are simpler Django tools, others have evolved to include React when greater interactivity was warranted.
Solum provides information to farmers to help them more optimally manage their cropping practices. One product to assist in precision agriculture is the Crop Advisor, an interactive web tool to help farmers and agronomists plan and budget the activities they will perform on their fields. This includes an interactive web Gantt chart, coupled with a weather-directed crop growth model and multitenant chemical and fertilizer catalog. I am managing a small team of web and software developers, including local full-time employees and remote contractors, to build this over approximately six months. I also designed the data model and developed the initial core functionality, including the core Gantt user interface and the crop growth model, which fetches weather data from NOAA servers.
Another of Solum's products is the No-Wait Nitrate system. This customer-deployed instrument analyzes soil for nitrate concentration, thereby informing subsequent fertilizer application. I architected and manage the development of an integrated system surrounding this instrument. Users scan bar-coded soil sample bags with an enterprise mobile app to geo-locate the sample, and then can see maps of their soil nitrate content and download PDF reports via a web portal.
Solum operates a highly-automated soil lab in Ames, Iowa. This lab leverages RFID technology to track soil samples as they flow through dozens of individual work stations comprised of both commercial instruments and proprietary robotics. These stations provide simple interfaces to lab operators and interact via REST to a unique cloud-based Laboratory Information Management System. Customers can then see their soil analysis results, such as maps and reports. Internal staff also utilize tools for automated and manual quality control. I performed initial technology validation, designed the system and data model, and led a small team of developers in its implementation, operation, and continued improvement.
3TIER provides Renewable Energy Information Services such as weather-driven energy forecasting and risk analysis. I led the software development function at 3TIER from 2007-2011, when we developed a number of innovative products for utility wind and solar plant developers and operators. These products are resident in a user-customizable, highly-available, and extensible Dashboard. For example, we built a Wind Power Forecasting product for wind farms. This event-driven system ingests real-time data feeds from both public sources and private meteorological masts, and applies neural nets and other sophisticated statistical models to generate forecasts from intra-hour to week-ahead time scales. I made key technical decisions, and hired and managed teams of developers and support personnel in the development and operation of this industry-leading product.
3TIER has considerable computing infrastructure, utilized to create global wind and solar data sets. These data sets required processing over 500 Terabytes of historical weather data to create over 100 Gigabytes of historical wind and solar conditions suitable for modeling renewable energy systems. I developed software to run on our 400 node (quad-Opteron) compute cluster, serving as a key developer on the global wind data set, and as the architect and manager of the system to compute the global solar dataset. I also designed and led the development of REST APIs and Interactive Web Tools to surface this unique data set, along with managing the development and operation of the considerable infrastructure software required to serve these data interactively.
At Amazon, I developed digital media products for millions of Amazon customers. With Search Inside the Book, I worked in the Media Technologies team to scale the digitized catalog to hundreds of thousands of books, and internationalized the interface for overseas locales. I also implemented an avant-garde 'popover' which shows information about a book when a user mouses over the book's cover. Among other features leveraging this valuable text corpus, Amazon provides a list of Citations derived from a book's text. I designed an algorithm to create a citation index and mine book text for citations, implemented this algorithm efficiently, constructed and managed a cluster of dozens of servers to process hundreds of Gigabytes of book text, adapted Amazon Distributed Hash Table technology to serve this citation corpus to hundreds of viewers per second, and finally implemented a simple web interface for these data. Both the popover and the citations functionality remain visible on one of Amazon's most visited pages, nearly 10 years after implementation.
At Amazon, I also served as the founding engineer for Amazon's Instant Digital Video category. This involved the transcoding of tens of thousands of high-resolution videos from content partners, and the careful normalization and ingestion of metadata for these movies and TV episodes. I led the creation of a curation tool to manage these video metadata. Launching a new product category at Amazon also involves significant project management in negotiating multiple organizations to integrate with Amazon's search, browse, catalog, navigation, payments, and other functions; I owned the creation of this category in Amazon's system and the necessary dependency resolution. I also interviewed over 100 developers to build out this team.