Cavity quantum electrodynamics encompasses the study and control of the interactions between quantum light sources and resonant modes of optical cavities. The subject of this thesis is cavity quantum electrodynamics with semiconductor quantum dots (QDs), which are light-emitting nanostructures with atom-like optical and electronic properties. Recently it has become possible to combine QDs with methods of producing cavities that have microscopically small volumes, which led to the observations of spontaneous emission enhancement, lasing, single-photon nonlinearities and vacuum Rabi splitting. These effects can potentially be exploited for applications in quantum communication, computing and metrology. A challenging obstacle faced in current research is the lack of control over the positions of the QDs within the cavity structures, because the majority of experiments employ self-assembled QDs that nucleate randomly at unpredictable locations during the crystal growth process. The technical objective of the present thesis was to address this problem by means of an alternative approach for QD growth that utilizes metal-organic chemical vapor deposition on GaAs substrates patterned with inverted pyramidal recesses. The QDs obtained by this approach are referred to as site-controlled pyramidal QDs, and the technique for their growth has been refined in our research group since more than a decade. One of the principal achievements of this thesis was to develop a deterministic and scalable fabrication procedure for integrating InGaAs/GaAs pyramidal QDs into planar photonic crystal (PhC) cavities. In particular, we succeeded in coupling single QDs and pairs of spatially separated QDs with a three-hole defect (L3-type) PhC cavities. Owing to the excellent site control and the few-meV inhomogeneous broadening of pyramidal QDs, we could routinely obtain a spatial alignment precision of better than 50 nm and thereby yield many effectively coupled QD-cavity devices on the same substrate. This facilitated systematic examinations of the coupling characteristics of single and pairs of QDs in cavities, without ambiguities related to the QD positions and the possible presence of spectator QDs in the cavity region. First, we investigated the L3 cavities containing a single QD at their centers in micro-photoluminescence and photon correlation measurements. We observed that the QD exciton line closest to the cavity mode becomes markedly enhanced iii Acknowledgements in intensity upon crossing the resonance through temperature tuning, which is a characteristic signature of the Purcell effect in the weak coupling regime. Furthermore, we performed detailed polarization-resolved studies and found that the QD exciton becomes co-polarized to the cavity only within a narrow detuning range. Most notably, we also discovered that pyramidal QDs detuned by more than 5 meV could not couple their emission to the cavity, such that the cavity resonance was spectrally absent or negligibly small even at high pumping power. This is in striking contrast with the typical behavior of self-assembled QDs, where a spurious "cavity feeding" mechanism contaminates the emission from the cavity with uncorrelated photons and leads to far-off resonance coupling. Using theoretical modeling of the optical spectra, we were able to understand the coupling characteristics of pyramidal QDs by taking into account that longitudinal-acoustic phonons can assist the excitation transfer from the QD to the cavity. A possible explanation for the absence of cavity feeding in pyramidal QDs is that the confined excitons do not interact efficiently with charges in the barrier material, unlike the situation in self-assembled QDs. Building on the knowledge acquired from our experiments with single QDs, we proceeded to systematically study L3 cavities in which 2 QDs were embedded with an interdot separation of 350 nm. Here the outstanding reproducibility of the excitonic states previously evidenced in the spectra from single pyramidal QDs turned out to be a crucial advantage, permitting the spectral identification of the individual QDs from the QD pairs. The most significant finding that emerged from our measurements was the observation of mutual Purcell enhancement from a QD pair, which constitutes the first demonstration of that kind. The results of this thesis demonstrate the benefits of site-controlled QD technology for cavity quantum electrodynamics and validate the potential of pyramidal QDs for implementing more complex architectures, such as multiple QDs in a cavity and nanophotonic integrated circuits consisting of waveguide-coupled cavities. A very interesting outlook regarding multiple QDs in a cavity is the exploration of collective effects like superradiance and multipartite entanglement. Further efforts in improving the quality factors of the cavities and reducing the linewidths of the pyramidal QDs may eventually culminate in reaching the coherent regime of strong coupling and achieving lasing.