Reshaping dataframe with column names inside the other column

Question

I have a dataframe with rather unintuitive structure:

device1    device2    column_name
      1          1    date
    111        211    param1
    112