For Eniola, the strongest angle is to leverage the TOPOLOGIX line, as it directly applies multimodal AI (ESM-2 protein language model embeddings + Morgan fingerprints) to a scientific problem (drug-resistance mutation prediction) with a clear benchmark and superior performance over structure-based baselines. Frame TOPOLOGIX as a foundation for 'scientific execution' by proposing to extend it to integrate multimodal data (e.g., sequence, structure, and experimental assay data) to predict resistance in real-world clinical contexts, aligning with the fellowship's focus on multimodal AI for scientific execution. Emphasize the ability to lead a focused project, the existing code and validation pipeline, and the potential to produce a publishable artifact, while noting the full-time Boston commitment is feasible given the M.Sc. program can be paused or taken remotely.

Full-time commitment in Boston may conflict with the M.Sc. Digital Health at HPI/Potsdam (Winter 2026/27), though the program allows graduate students to take leave; remote is by exception. The program's focus on physical laboratory systems and multimodal data is less central to Eniola's core neuroscience and dynamical-systems work, so the proposal must clearly bridge to scientific execution. Also, Eniola is an independent researcher without a PhD, but eligibility includes graduate students, so this is acceptable if the M.Sc. enrollment is leveraged.