Did something like that at Nijmegen in the late 1990's. The lasers involved worked at wavelengths too low for PMT's so I stuck a wideband variable gai n amplifier onto a photo-diode
Stuck two fast Burr-Brown VGAs in series to get fast forward gain, and two cheap slow Linear Technology parts in series to do the DC feeback to keep t he average long-term output at 0V.
Decoupling the rails well enough to keep the whole mess stable was demandin g, but I designed it onto a printed circuit board in one hit.
I used a Texas Isntruments CMOS-input op-amp to monitor the average output, and the data sheet failed to mention that it had an input capacitance of a bout 15pF which I hadn't compensated, so the board oscillated (rather slowl y) when I first turned it on. A 3.3pF compensation capacitor fixed that, bu t stuck out like a sore thumb on surface mount board.
I hadn't been fond of Texas Instrument before that , and that didn't make m e any fonder.