I was brought in to Two Sigma to help them design an AI powered air hockey table.
Two Sigma puts a lot of faith on their team to know more about emerging technology than their competitors. To keep their skills sharp, they have regular challenges to get employees collaborating on projects that hinge on new tech. The challenge is streamlining the development process so they maximize time learning and experiment, and not getting frustrated with the constraints imposed by x-factors like the precision of the machine tools available at the internal hackerspace.
I came in as a mechanical and DFM consultant to figure out a methodology to iteratively prototype the physical hardware necessary to get a fast, accurate robotic air hockey table running reliably enough to concentrate employee time on developing the AI on top of that hardware. I worked with Trammell Hudson to transition the prototype they’d constructed into a wire-control system similar to that used in the “Bat-Cam“. This allowed the team to prototype subsequent iterations of the system using off-the-shelf FDM printers and standard motors rather than custom gantries.