I'd have a host of BJTs in TO-92 packages, but not that exact one. I'd guess that a lot would work, but they are using it for an amplifier on a very small signal so some ranges of alpha gain and I-V characteristics may not get you what you're looking for. AC coupling into a BJT base can get a little interesting to tune/ understand, since the BJT acts as a non-linear voltage based current sink (base to emitter), so band pass filtering on that node could be pretty dependent on BJT characteristics . . . I'm no expert here, BJTs are old school stuff.
I took a quick read over that page and that circuit isn't doing anything crazy, and it's using crazy old tech. They're using an actual bjt amplifier, used according to them because the op amps will cause too much signal distortion . . . then they grab a cheap, 30 year old op amp with notoriously terrible characteristics. Yea, of course it's going to cause too much signal distortion. If you just design this circuit with a good chopper-stabilized auto zeroing high gain amplifier, you can get <10uV of offset across temp and <1pA normal input leakage with maybe a few hundred across temp . . . makes that whole circuit kind of pointless. They're using an opamp for a comparator too, which tells you how little they care about high frequency stuff, they really just need high gain on a low amplitude signal, which is pretty easy to do these days.
Also worth pointing out these days you can get high speed ADCs that could actually natively read signals in that kind of amplitude range at high enough speeds to catch the waveforms you're looking at. Sounds like you only care about pulse counting (frequency measurements), so this is probably overkill and would require extra software.
I may not know your exact use case, just my 2 cents based on 2 minutes of quick skimming,
Kyle