During these last decades, by virtue of observations, the Standard Cosmological Model has emerged, providing a description of the Universe's evolution using a minimal set of independent constraints - the cosmological parameters. Among them is the expansion rate of the Universe, the so-called Hubble constant or H0, first measured by Lemaître in 1927. The century that followed this cornerstone measurement saw numerous attempts to refine the initial value, and for good reason: a precise and independent measurement of H0 will bring strong constraints on the cosmological models. It could notably help the astronomers to better understand the nature of dark energy, thus making it one of the most sought-after prizes in modern cosmology. My work at the Laboratory of Astrophysics of EPFL is embedded in this context. I am part of the COSMOGRAIL and H0LiCOW collaborations, aiming to measure the Hubble constant with the highest level of precision using time-delay cosmography, a method based on the theory of strong gravitational lensing. This effect occurs when an observer looks at a light source located behind a massive foreground galaxy. The mass of the galaxy acts similarly to an optical lens and focuses the light rays emitted by the source. As a consequence, multiple lensed images of the source appear around the lens galaxy. If the luminosity of the source changes over time, the variations will be seen in all the lensed images but with a temporal delay due to the different travel paths of the light rays. By carefully monitoring the luminosity variations of each lensed image, one can precisely measure the temporal delays between them. Combined to high-resolution observations of the foreground galaxy and its surroundings, it is possible to directly measure the Hubble constant upon the sole assumption that the General Relativity is correct. Since more than 13 years, COSMOGRAIL monitors dozens of lensed quasars to produce high-quality light curves and time-delay measurements. During these last four years, I took care of the monitoring schedule, continuous data reduction and time-delay measurements through the development of curve-shifting techniques. I produced light curves and measured time delays on a variety of lenses. After more than a decade of endeavours, COSMOGRAIL and H0LiCOW finally revealed their measurement of the expansion rate of the Universe from a blind analysis of three lensed sources. I had the privilege to be the lead author of the publication presenting our measurement of the Hubble constant, H0=71.9 -3.0+2.4 km/s/Mpc 3.8% precision in the Standard Cosmological Model. Such a precision allows a direct comparison with the results of the distance ladder technique in the local Universe and the Planck satellite Cosmic Microwave Background observations in the distant Universe, both of which being currently in a significant tension of unknown source.