This document summarizes Makoto Kuwata's presentation on creating a high-speed template engine in Python. It discusses Tenjin, a template engine created by Kuwata that is very fast. It benchmarks Tenjin against other template engines like Mako, Jinja2, and Django and shows that Tenjin is significantly faster. The presentation provides optimization techniques for template engines like using bound methods, local variables, and format strings instead of string concatenation. It also recommends implementing performance critical parts in a low-level language like C to achieve better speeds.