Effectively enabling chemists to utilize upcoming exascale resources requires more than massively parallel chemistry codes: the tools used to perform analysis and generate results must also benefit from these parallel developments, particularly to enable in-situ analysis that can both greatly shorten time to solution and enable novel insights. A serial graph analytical tool, ChemNetwork, has been rewritten as a package usable in a massively parallel molecular dynamic simulator (LAMMPS) in order to enable in situ analysis taking advantage of domain decomposition and data structures within LAMMP S. Key functions within ChemNetwork has been reimplemented in parallel utilizing MPI and MPI I/O to yield a negligible performance overhead when run in parallel within LAMMPS. LAMMPS was also used to help understanding and designing new models for Li ion battery electrolyte materials. In this project, a workflow was implemented to sample pressure and temperature phase space for typical electrolyte materials. In another collaborative project, hyperparameter optimization was performed to determine the optimal method for running deep learning simulations for ECP-CANDLE benchmarks, achieving a 10x speed up. Finally, I will discuss how my experience developing programs, mentoring, and using leadership class computing resources will help me in developing workshop and training programs for the ALCF.