There are compelling motivations for the existence of dark matter particles which would explain the invisible gravitational mass present throughout the universe. Interest in the field has increasingly turned to searches for lighter dark matter candidates with mass < 10GeV/c^2, which deposit less energy in the detector and are therefore difficult to see with most detectors. The Dark Matter in CCDs (DAMIC) experiment has advanced the field of such light dark matter searches through the use of silicon charge-coupled devices (CCDs). CCDs combine high spatial resolution and low noise to provide a unique detector for multiple candidates for dark matter. The current generation of the experiment located at SNOLAB has been used to look for several dark matter candidates. The next generation, DAMIC-M at the Modane Underground Laboratory in Modane, France will utilize a large array of new skipper CCDs, capable of achieving sub-electron noise resolution by using multiple non-destructive charge measurements, to search for multiple dark matter candidates. The first part of this thesis describes the development of new high performance low noise electronics for CCD readout and control, the Online Digital Interface for Low Noise Electronics or ``ODILE'' system. We first describe the firmware used for these ODILE controllers, and follow this with a series of tests integrating the ODILE development with skipper CCDs to achieve noise of <<1 e- RMS/pixel. Use of these ODILE controllers will allow reliable operation of the O(100) CCDs necessary for DAMIC-M, running these CCDs in a steady state condition for months at a time to conduct dark matter searches. In the second part of this thesis, we describe the development and validation of Monte Carlo simulations of the DAMIC at SNOLAB detector, which have been used to build an extensive model of its dark matter search backgrounds. This model was used to set limits on weakly interacting massive particle (``WIMP'') dark matter cross sections in the 1-10 GeV/c^2 mass region; an excess over the known sources of background was found, well-parameterized by an exponential with characteristic energy 67 +/- 37 eV. This result will help guide the development of the DAMIC-M detector, through a betterunderstanding of our existing backgrounds and providing an interesting unknown excess to explore. Finally, we describe a novel measurement of the rate of activation of tritium in silicon by cosmic ray neutron spallation. This measurement used a neutron beam at the LANSCE ICEHOUSE facility to irradiate silicon CCDs. The resulting tritium activation in these CCDs was converted into a measurement of the sea-level cosmic-ray neutron activation of silicon: 112 +/- 24 atoms/kg/day. This is the first time such activation has been measured in silicon, and sets importantconstraints on the exposure to surface cosmic rays that any silicon-based dark matter detector can have. With this result, DAMIC-M will be able to meet its tritium background budget of 0.1 event/kg/\kev/day in our region of interest.