Understanding how the turbulent motion of plasmas (ionized gases) leads to the generation of magnetic fields, and in turn, how the energy stored in these magnetic fields fuels particle energization are grand challenges on the forefront of space and astrophysics. Applications range from space weather modeling in the near-Earth space environment, to magnetic field at the largest scales of the universe. Addressing these phenomena requires self-consistently evolving the plasma constituent particles in space and time, with the electromagnetic fields they generate; this represents outstanding numerical challenges: 1. These are multiscale phenomena requiring a spatial resolution bridging between the large scales of bulk flows and the small scales of the relevant particle dynamics. 2. They are essentially kinetic: wave-particle interactions and strong deviations from thermodynamic equilibrium are essential. Thus, adequate modeling approaches either resolve these processes in phase space or they employ effective models for the unresolved velocity space behavior. Particle-in-cell simulation codes are in the former category, representing the highest numerical expense and highest physics fidelity. Advanced fluid modeling takes the latter approach, allowing three-dimensional truly multi-scale simulations at reasonable computational expense. In this project we will take advantage of both of these approaches. This project will focus on three phenomena: (1.) Electron energization in reconnection in the near-Earth space environment. Reconnection is a process where magnetic topology change enables the release of energy stored in the magnetic field, and it is believed to be responsible for the most extreme particle energization processes in the universe. To understand the rate of energy conversion and its partition, we will analyze the multi-scale interaction between particles and electromagnetic fluctuations due to instabilities. (2.) Astrophysical dynamo converts energy from turbulent flows to magnetic energy and it is responsible for the magnetization of the universe on the largest scales. This is an inherently three-dimensional phenomenon, which has almost exclusively been studied with magnetohydrodynamics, a description that does not retain sufficient complexity of the phase space dynamics. Going beyond the state-of-the-art, we will use advanced collisionless fluid models, enabling us to effectively account for crucial micro-physical processes. (3.) Ion streaming instabilities in magnetized plasma shocks can cause ion heating and thus may play a key role in the long-standing injection problem of cosmic ray acceleration. The magnetosphere of the Earth is an ideal laboratory, where multi-spacecraft missions allow an in-situ observation of kinetic phenomena, with phase-space-resolved measurements. Such ion instabilities have recently been observed in the bow-shock of the Earth, and we will provide theoretical support to these observations through a numerical analysis of the phenomenon. Topics (1) and (3) is studied within the Wallenberg project “Extreme Plasma Flares” aiming to establish the conditions for the most extreme particle energization processes, combining space observations and advanced computing. Topic (2.) is studied within the VR project “Data-driven optimal models for kinetic dynamos”, using data-driven model identification to construct advanced fluid models for dynamo modeling.