Treating immunotherapy-refractory tumors faces clinical challenges like resistance driven by tumor heterogeneity. Industrial hurdles include the need for innovative engineering of therapeutic modalities. Advancing new treatments requires novel targets, and beyond state-of-the-art translational routes. This thought process is particularly applicable to targeting macrophages in immunotherapy-resistant tumors. Previous clinical and industrial efforts to deplete macrophages for eradication of such tumors have largely failed. It is now well-accepted that reprogramming the macrophages might be the best way forward however this remains difficult due to their plasticity and heterogeneity. Overcoming these obstacles requires precise macrophage reprogramming, novel antibody engineering, and a deeper understanding of clinical myeloid context. During my presentation, I will discuss how we are currently tackling these challenges using latest multi-omics/spatial approaches and AI tools that are being constantly improved through functional lab-in-the-loop R&D cycles – the basis of our beyond state-of-the-art RETROFIT platform at KU Leuven, Belgium. This platform has already revealed a series of innovative myeloid/lymphoid targets and at least one best-in-class multi-specific antibody capable of eradicating preclinical tumors resistant against all previous (T cell/myeloid targeting mono-specifics) and current [PD(L)1 x VEGF bispecific] generations of immunotherapies.