Current observations and models indicate that supernova feedback processes are important in shaping many properties of galaxies. In spite of rapid progress, our current knowledge of the physical properties of super-winds is still in its infancy: it comes mainly from few nearby galaxies or from bright ultra-luminous infra-red galaxies where the kinematics of the winds are inferred from the blue-shifted sodium absorption of the cold entrained gas against the galaxy continuum, i.e. along radial sight lines. We are studying low-mass galaxies "caught in the act" of producing super-winds half-way across the universe (at z=1) using the MgII resonance line in QSO spectra. A unique aspect of our project is that we are combining the 2D spatially-resolved kinematics of starbursts with the 1D kinematics of the absorbing gas seen along the quasar sight line. Our scientific goals, however, extend much beyond our sample. While at the outgoing host, the University of California, Santa Barbara (UCSB), working with Prof. C. Martin, we aim to combine our results on intervening sight-lines with upcoming surveys on radial sight-lines in order to put both in a unified context. This will give us constraints on the radial properties of outflows. In addition, given that our sample nicely fills unexplored mass and SFR ranges, we will quantify how the wind properties vary over a much wider range of masses and star-formation rates. At the returning host, the Laboratoire Astrophysique de Toulouse-Tarbes (LATT), the fellow will prepare, design and exploit the next-generation of feedback surveys using the wide field integral field unit, MUSE, which will provide unprecedented 3-dimensional data sets on thousands of faint emission line galaxies, and spectra of hundreds of galaxy-galaxy and galaxy-quasar pairs. In other words, the skills acquired at the outgoing host will be very valuable to exploit the MUSE instrument, which is one major training objective.