Closing two chapters
I'm leaving TTT Studios after seven and a half years and Amanda AI after three. What I learned growing an engineering team from ten to fifty, and co-founding a company through a pandemic.
This is my last week at both TTT Studios and Amanda AI. I’m starting a new job next month, which I’ll write about separately. Before that I want to write down what I’m taking from the last several years, mostly for myself.
I joined TTT in October 2013 while I was still finishing my master’s. There were about ten engineers. By the time I became VP of Engineering we had more than fifty, across five engineering managers, building apps and systems for clients in over fifteen industries, from energy utilities to medical records. For most of that time I was learning management by doing it wrong first.
The biggest thing I learned there is that an engineering organization is a system you design, like any other system. Hiring, onboarding, how work gets assigned, how people grow, what “done” means, who decides technical standards. When those are unclear, good engineers still produce bad outcomes, and you spend your time fixing symptoms. When they’re clear, most things take care of themselves. The work that mattered most from me wasn’t the code. It was the hiring loop, the career ladder, the review process, and above all the managers we developed. Five good managers do more than any number of great individual engineers.
The second thing is from client work. Across dozens of projects, the technology was almost never the reason something failed. Unclear ownership, the wrong problem, a user nobody had talked to. The best technical discovery sessions I ran were mostly questions about how the client’s business actually worked.
Amanda was the harder chapter, and the one I learned the most from. We built on-device face recognition for conference check-in, raised more than ten million dollars, and cut attendee wait times by as much as 80 percent at the events we ran. Then the pandemic erased our market for a year. I learned what it means to own a technology strategy and a budget at the same time, how to pitch investors, how to say no to money that came with the wrong use case attached, and how to keep a small team going when the thing you built doesn’t have anywhere to go for a while. I’m proud that we never built the surveillance products people kept asking us for.
I’m also leaving Vancouver, the city I’ve lived in for nearly ten years since I came from Wuhan for grad school. That part is harder to write about.
What I’m taking with me: a strong belief that product, engineering and organization design are one job, not three. A healthy skepticism of any model’s benchmark numbers until I’ve seen it in the real world. And a lot of gratitude to the people at both companies, who put up with me learning in public.