I've done a few bits of design work for satellites, so I'll give a brief overview of the design process.
An initial design will be done and carefully analyzed and any interesting areas will have a breadboard or prototype made.. Nothing different there from a commercial project.
Then the real; fun starts with the documentation. Design description, which would include everything from a general overview to the calculations used to determine the value of every component on the board, simulations, results from the breadboards and explanations for any differences. Then there's the Compliance Matrix, worse case analysis, parts stress analysis, radiation analysis, MTBF calculations and so on. It's been a while and I'm sure I've missed out a few of the design documents. Further documents will include plans for further stages, so a test plan for an amplifier would say that it will be tested at nominal, max and min supply voltages, and at various levels and frequencies with an assortment of specified waveforms and at what temperatures.
At that point, it's time for the first formal design review, with the customer, possibly the end customer and also critical suppliers.
About the same time, there's a whole pile of process and parts work happening. Each process needs to be space qualified. If you don't have an up to date certificate to say you can hand solder 0402 resistors with Sn53, and another with a certificate for inspecting the same then it doesn't happen. Before the iron is even turned on the documentation will need to specify solder and flux types. If the board assembly is on a P^P and reflow then that line will need space qualifying.
LAT testing, or Lot Approval Testing will be going on. For each reel or tube of components a few will be taken and subjected to an accelerated life test. If one component from that lot fails, the whole lot is binned and a new lot tested. The accelerated life tests typically take 12 weeks, so a failure means a three month delay on the project. Ideally you pick what parts you need early on and get the LAT testing done early. Sometimes you need to change a resistor value, so the wise but expensive move is to LAT test more components than you need.
Before production of the flight hardware starts then the second big design review happens. All the previously mentioned documentation is updated, usually by this stage a proper prototype that is intended to meet spec has been made and electrically tested with some temperature testing added in, along with any other testing that is required by the customer to provide confidence. Additional documentation would include the Test procedure, which would take the Test plan and fill in exactly what measurements are to be made, what equipment is to be used and test limits.
Only after that design review will production of the flight hardware proceed.
The testing of parts that are going into space is tedious. On a commercial product, designs are tested at ambient and at temperature extremes. For a space project, they are tested over a wider temperature range every 10 or 20 degrees C and in vacuum, in a Thermal Vacuum (TVAC) chamber, Usually the equipment needs to be watched whilst it is in the chamber, so the three hour wait for the temperature to stabilize is tedious, I've read a lot of books in the wait time, and ate a lot of snacks and put on 20kgs.
The vibration testing is fun. Launch vibration is so high that no-one is allowed in the same room at the vibration table when it's operating and from the next room it actually sounds like a rocket taking off.
Finally all the test results, and process verification documentation is collated and a final design review is carried out. A typical module in a satellite will generate about 10m high of paperwork. There are a lot of modules in a small satellite and there's more paperwork involved as the modules are integrated and tested.
What I like about doing design work for space applications is that you don't get caught out at the end of a design due to skimping on analysis early on. You never need to do a quick board respin as a resistor is dissipating more power than previously anticipated. Designed get done properly. What I don't like is the huge amount of testing required. When pen plotters where the normal way to record a result, I know of someone who had done 9000 of the required 11000 plots when he was told that the project had been cancelled. Fortunately GPIB measuring is now the norm.