After my Ph.D. and postdoc, despite having a great experience, I was ready to be done with academia for a while. It was also difficult to accept one of the academic positions on offer in New England because I was waiting for Cydney to be done her Ph.D. and move back to Vancouver. I had met Ian Davis during my time at M.I.T. and started working for his game company Mad Doc Software as an AI programmer, and later as the studio AI head. I worked on a never-to-be-published game that got cancelled by the publisher and ended up running a small team that contributed boss AI to the Ironman movie game (don’t play it, it’s terrible - I like to think not as a fault of my team) as well as Brothers in Arms: Hell’s Highway. In 2008 Rockstar Games acquired Mad Doc Software and turned it into Rockstar New England. I ended up working a bit on Red Dead Redemption’s animal AI after the purchase, but as Cydney was done her Ph.D. and had a job offer in Vancouver it took the support of Ian Davis and a phone call to Rockstar Vancouver to be transferred there. With Ian’s support I did get to publish a paper on some of the AI planning work I produced there and later at Rockstar Vancouver - though Rockstar supported this so little that I had to sneak out and give a talk about it with Damian Isla without mentioning Rockstar at GDC 2009.
In Vancouver I worked on Max Payne 3 almost exclusively until 2012, for better or for worse. I cannot argue with Rockstar’s success, but its attitude towards employees leaves much to be desired. For years our project was on a backburner due to Rockstar’s cash cow Grand Theft Auto plodding towards its 4th installment. Internal politics seemingly out of the sight of the company’s leaders whiplashed us with direction and leadership changes. We went from an open world game in slow increments towards one of the most linear titles published. Alongside, my AI work changed from exciting possibilities to supporting mostly scripted and staged scenes. When we could get the attention of leadership, it meant crunches to finish some random demo that nobody new they even played in the end. Employees were clearly deemed utterly replaceable unless they put in stupid hours until 6am for no good reason, checking in all sorts of ludicrous bugs that we had to fix later. I thought that working for a company owned by a publisher would be better than being at the whim of an external publisher, but that ended up being mostly untrue. Half way through the project Rockstar threw out all the code we had written and made us rewrite everything using Grand Theft Auto’s engine. The subtle bugs this caused are still in the shipped game (I was fixing one the night before shipping, and the fix did not make it into the shipped copy), and a signifcant percentage of great engineers left or were fired in the aftermath of this. I probably should have left, too, but I told myself I would see a game through to release, and I did. While I miss working on games professionaly, I am very happy I got out of that type of game industry.