Today is my last day at eBay. I have been there for nearly four years and it has been a remarkable experience. When I joined eBay, I had come from a startup called Ludi Labs. One thing I realized while at Ludi Labs was that "scalability" was largely a hypothetical term to me. I thought I understood the keys to making an application function at web scale, but it unfortunately it was very theoretical knowledge. None of the startups I had worked at had ever really hit that kind of scale. Hence my interest in working a large scale web company, like eBay.
Four years later and I realize how very limited my theoretical knowledge of scalability really was. Sure I understood the basic principles, but there are definitely many subtleties involved in implementing these principles. I won't now claim to be a scalability expert, but I have certainly learned some real world lessons.
When I first joined eBay, I worked with the infrastructure team on eBay's homegrown presentation technology. A lot of this technology was based on what works at scale: high volume web pages, working in dozens of different languages, running on servers spread across multiple data centers, and built by hundreds of engineers around the globe. I went on to help many of the different teams at eBay adopt and optimize their use of eBay's presentation technology. As such I got opportunities to work on many of the major pages of eBay: home page, search results, view item, my eBay, just to name a few. Often my experience was limited to helping the team solve some site speed issues, but I still got a chance to understand how each of these highly complex web pages could function at massive scale.
This leads to the obvious question of why I am leaving eBay. The future for mobile at eBay is very bright. However it's one of those cases where the change is not about where I'm coming from, but about where I'm going. In this case that's Bump Technologies. I'm very excited about the opportunity at Bump. They already have a great product that they've built in just two years. However, I think that two years from now the current product will look very limited in comparison. It's going to be an exciting time.