Petri nets based testing

What to do when tests cannot catch bugs which occur because of weird states of a system under test, known as "edge cases"? Is it possible to increase test coverage and find more issue without creating excessive tests or sacrificing their execution time?

In this talk, we’ll discuss how we faced this problem at Toptal, started moving from usual tests to model-based testing, what problems we faced during this transition, why we use Petri net instead of state and what we ended up with. The talk is illustrated with examples of Petri nets and plenty of Ruby code.