I am doing some manipulation on a data frame:

df

  Node        Interface      Speed      carrier    1-May  9-May   2-Jun    21-Jun  
  Server1      internet1     10          ATT       20     30     50      90    
  Server1      wan3.0        20          Comcast   NaN    NaN    NaN     100
  Server1      wan3.0        50          Comcast   30     40     40      NaN
  Server2      wan2          100         Sprint    90     70     NaN     NaN
  Server2      wan2          20          Sprint    NaN    NaN    88      70
  Server2      Internet2     40          Verizon   10     60     90      70

I need to merge rows in data frame group by Node and Interface, replace nan values with the other row and pick the max value for the speed for the interface.

expected data frame should be like this:

df1

   Node        Interface      Speed      carrier    1-May  9-May   2-Jun    21-Jun  
  Server1      internet1     10          ATT       20      30      50       90    
  Server1      wan3.0        50          Comcast   30      40      40       100
  Server2      wan2          100         Sprint    90      70      88       70
  Server2      Internet2     40          Verizon   10      60      90       70

I tried this:

df2=df.groupby(['Node','Interface','carrier']),agg({'Speep': 'max'}).reset_index()

df3=df.drop('Speed', axis=1)

df4=df3.ffill().drop_duplicates()

Not quite working. Is there an easy way to merge rows, replace nan values with the other row values and pick the max Speed for the Speed cell value?