Laser-induced incandescence (LII) is a well-established technique for tracking soot, potentially enabling soot volume fraction determination. To obtain crank angle resolved data from a single cycle, a multi-kHz system should be applied. Such an approach, however, imposes certain challenges in terms of application and interpretation. The present work intends to apply such a high-speed system to an optically-accessible, compression ignition engine. Possible problems with sublimation, local gas heating or other multishot effects have been studied on an atmospheric co-flow burner prior to the engine experiments. It was found that, in this flame, fluences around 0.1 J/cm2 provide the best balance between signal-tobackground ratio, and soot sublimation. This fluence is well below the plateau regime of LII, which poses additional problems with interpretation of the signal. This is especially true when a wide span of temperatures and gradients is present, as encountered in diesel combustion. The purpose of this study is to provide an initial investigation into the feasibility of high-speed LII in engine applications. Particular attention will be given to the late combustion phase, i.e. after the end-of-injection, where phase-averaged data might not provide sufficient information. The pitfalls related to engine data interpretation are briefly discussed and some qualitative results are shown, which have been verified with high-speed luminosity measurements. To apply high-speed LII in an engine one should take utmost care with the selection of detection equipment (e.g. filters and detection gate) and settings to suppress the strong background radiation. All in all, the technique is found to be feasible, but the balance between fluence, signal strength, and effects on soot morphology was shown to be extremely delicate, and proper detection equipment is a requirement.