The Large Hadron Collider (LHC) is delivering the highest energy proton-proton collisions ever recorded in the laboratory, permitting a detailed exploration of elementary particle physics at the highest energy frontier. It is uniquely positioned to detect and measure the rare phenomena that can shape our knowledge of new interactions and possibly resolve the present tensions of the Standard Model. LHC experiments have already observed the long-sought after Higgs boson and have achieved unprecedented levels of sensitivity to new particles at the TeV scale with on-going searches for new physics, including dark matter. This trend is expected to continue during the next LHC run and with the High-Luminosity Large Hadron Collider (HL-HLC), anticipated to start data taking in 2027. New ideas for event reconstruction and data analysis are required to address the experimental challenges posed by the complex experimental environment at the HL-LHC that arises from a significant increase in pile-up, or extra particle collisions of protons traveling in the same bunch, leading to far more complicated event signatures at the HL-LHC. In my talk, I will discuss the application of state-of-the-art machine learning methods, including graph neural networks, to new physics searches at the LHC, detector reconstruction, event simulation and real-time event filtering at the LHC. I will also discuss related cross-over machine learning applications to searches for dark matter substructure with strong gravitational lensing with the upcoming Vera Rubin Observatory.