Also creating a vein on the leading edge of the guide on the roof of the port to help steer the airflow around the guide helps also. You might experiment with Bondo first to see about size and shape. But I would have the leading edge more blunted than sharp. After flow testing with Bondo, I would build-up with some brass brazing rods to make something more permanent, and do the same for the "lumps" on the floor of the port.
The more I look at those valve guides the bigger interference they seem to be. Once the ultimate flow geometry around the guide is determined why not make guides that incorporate those aerodynamic features and just orient and press them in?
