Here is a design for a no feedback MC amp with a gain of 15. It utilizes paralleled low noise JFETs in a diffamp configuration. I purposely avoided feedback, primarily because the feedback chain introduces resistor-based thermal noise. Also, for the small signal swings of tens of mV it's possible to get good linearity w/o feedback. Sim results look good, with distortion <-100 dB and a 0.4 nV/sqrt(Hz) noise floor. The circuit has been simulated over 0-75 Centrigrade and appears well behaved. I have not started running a parametric sensitivity test. Perhaps some one who has designed similar circuits could indicate which transistor parameters need to be swept.
