At Athyria, we provide software development services to fortune 100 leaders in aerospace, agriculture and retail. My role is both as an individual contributor and as a founder (sales, marketing, coffee fetching, etc.). As an individual contributor, I am currently the chief architect for Boeing's API strategy. If you're not sure what this means, be sure to read Steve Yegge's Rant linked below (the photo is not of Steve, btw). I worked directly for Steve for several years at Amazon creating Bezos' dream of APIs everywhere. Now, I am part of the Amazon diaspora spreading that dream throughout the world.

Harbor Vascular is a provider of medical services. Currently in the development phase of our first service, we have a working medical telemetry device, data from human trials supporting the efficacy of the device and a patent filing. More on the service in the near future. As for now, it is under wraps.

I was the first software developer hired at Base 2. I hired and built out project teams for deployment at clients. I also led teams providing architectural design, team leadership and software development. I focused primarily on aerospace, but also did some work in heavy industry. We built out CI/CD for our software with full automated tests from unit tests through system testing. I helped implement our AS9100 Quality Management certification where we passed each of our audits.

Design, develop and team-lead the development of an air-to-ground telemetry system for capturing "useful" airplane telemetry. This "useful" metric is programmable and directs which data are delivered to ground and and when those data are delivered (saving data transmission costs for non-critical data). This work lead to US Patent application number 20160110368.

Designed and developed RESTful web services using Java with Spring and JAX-RS/CXF frameworks, Memcached, MongoDB and other current technologies including an API suite for a multi-client video streaming application.

Have you ever wondered what might be fun things to do in your area right now? Perhaps you are hosting an event and want to get the word out? ToDoFinder can help! (er...could have helped). As a one-man company, I developed a "geo-temporal" search engine to facilitate discovering events occurring at a specific location and time. I filed a patent to protect my invention, but was beaten to the punch by just over two years.This was before Apple released their MapKit, so I developed my own MapKit borrowing map tiles from Google and/or Microsoft. ToDoFinder's MapKit was very responsive and featured dropped pins and text bubbles much like Apple's own Map Application.In the end, ToDoFinder failed for a number of reasons. First and foremost was a problem I confirmed with Tim O'Reilly of O'Reilly and Associates: I had no realistic plan to achieve critical mass of content and users. Content attracts users...which attract more content. No content attracts no users...which attract no content. Realizing this after some time, I switched gears to simply license my MapKit...which lead to the second reason for ultimate failure: Apple released MapKit. Apple learned of my work and flew me to Cupertino. Apple executives apologized to me personally, which I think is pretty cool even though they squashed me like a bug. It happens. We learn. We move on.

You can't manage what you don't measure, so we adopted one of Amazon's very first "two pizza team" metrics: reduce the number of customer contacts per order (later revised to number of contacts per dollar). We first developed the means to measure this, then executed toward the stated goal to great success. I managed the Customer Service portion of a number of efforts during this time including Amazon Prime, Return Authorizations, DVD Rentals in EU, Harry Potter Delivery on Publish Date, Multiple Accounts Same Email, and many, many more.In 2005, I was hired as the first TPM for the Marks and Spencer project. M&S is a large UK retailer who needed not only a web presence, but also an in-store tool for their sales associates and an ordering tool for their phone customer service agents. I was in charge of both the in-store tool and the phone tool. Amazon had never built an agent assisted ordering tool before. I traveled to London several times to meet with the client, work in their stores to understand the experience and help write requirements. When deploying to over 800 stores I recommended a "rolling thunder" approach rather than a "big bang" approach. This allowed us to validate the technology early with a small set of employees and also allowed us to train thousands of employees without significant operational disruption.M&S, like other websites we had built (e.g. Target, Gap, Toys R Us, Bebe), was built by identifying an existing site similar to our goal and branching from that code, an untenable approach. I spearheaded an effort to componentize the functionality of our websites starting with the Order PipeLine (OPL). The OPL we were replacing was 51k lines of code. We winnowed that down to 1500 lines and just a few hundred lines of configuration. Not only does this reduce maintenance costs, but we measured that implementing a new OPL with our technology reduced the effort from 2 years to 4 weeks. We repeated this success with detail pages and others.

Amazon recruited me in early 1999 into a ten-member team charged with ensuring Amazon.com would not be crushed by customer demand in the 4th quarter of 1999. At the time, Amazon was built on a single database, ACB a.k.a. Amazon.com Books, and all data access whether from the website, customer service, fulfillment, supply chain, all of it hit ACB directly. We quickly built services to act as a middle tier to conserve connections to ACB, then began splitting the database into separate components each with their own service front-ends: Customer Master, Item Master, Order Master, etc. This was our first step toward scaling to the truly enormous.In 2000, I joined my primary client, Customer Service Software working under Steve Yegge where we built a new tool, Arizona, for Customer Service Agents to handle customer requests whether over the phone or through website contacts. Once Arizona was up and running, I transitioned to being a Technical Project Manager (TPM) for the Arizona team.

Led the technical development of an enterprise application to stream and track video advertising for the cable television industry, from design, implementation, and field testing.

Having just seen NCSA Mosaic in March of '93 (Mosaic was the first browser and was released in late January of '93), I worked with TIG founder Michael Bauer as the tech guy to build out our first web presence hosted initially on a machine 'borrowing' a net presence from his daytime employer. For a few months, TIG was just Michael and me. Michael had landed a consulting gig for O'Reilly & Associates' GNN, a web-based tech magazine, where we worked on their concept of selling web advertising. In September of '93, GNN became the first company ever to sell an Internet advertisement. I do remember reviewing the GNN business plan in the summer of '93 and offering that these advertisements could include images rather than simply linked text. I suspect I was not the first to suggest this, but it is possible that I was.

Transarc was a Carnegie Mellon / IBM startup founded in 1989. We had two primary products: distributed file systems, AFS and DFS, and a transaction processing monitor, Encina. This monitor enabled its users to build transaction aware applications across networks. I helped develop Encina, being the primary developer for the transactional B-Tree, one of the persistence models in our CICS offering. On the side, in 1993, I wrote scripts to convert our printable SGML documents to a new format called HTML. I demoed this new technology to our executives and proposed that we integrate with HTTP servers to enable web-based transactions (e.g. e-commerce). My recommendation fell on deaf ears, though years later would see Encina included in IBM's WebSphere. Frustrated, I helped start The Internet Group in 1993 while staying at Transarc one more year.