MOTIVATION LETTER The AIMS model of mathematical training, delivered by African faculty on African soil, is the missing structural piece in my research trajectory. I am a Nigerian pharmacist and computational researcher with a B.Pharm from the University of Ibadan (CGPA 5.1/7.0, German equivalent 1.9) and current enrollment in the M.Sc. Digital Health program at the Hasso Plattner Institute, University of Potsdam. My research applies Bayesian MCMC, dynamical systems, and topological data analysis to biomedical problems, from addiction neuroscience to drug-resistance prediction. What I lack is the deep, formal mathematical grounding that AIMS provides, and what AIMS gains from me is a researcher who has already deployed advanced mathematics on real clinical data and will return that capacity to African institutions. My current project, TOPOLOGIX, illustrates the fit. I use ESM-2 protein language model delta-embeddings combined with Morgan/ECFP drug fingerprints and a Random Forest classifier to predict drug-resistance mutations from sequence alone. The model achieves AUROC 0.804 plus or minus 0.025 on the Platinum benchmark of 553 mutations, and 0.634 on SKEMPI 2.0. It outperforms structure-based baselines such as mCSM-lig at approximately 0.70 while covering 100 percent of mutations, compared to roughly 18 percent for structure-limited tools. This project sits at the intersection of machine learning, protein biophysics, and clinical pharmacology. The mathematics underneath it, representation learning on biological sequences, uncertainty quantification, and classifier calibration, is exactly the curriculum AIMS teaches. My prior work includes a pre-registered, powered replication study on hERG cardiotoxicity using bipartite persistent homology, which found that topological features do not beat a plain descriptor baseline (AUROC 0.8426 versus 0.8782). That negative result, reported honestly, shaped my pivot to sequence-based methods. AIMS training would let me formalize the statistical and algebraic foundations of that pivot. My commitment to African scientific development is documented in my work history. I worked as a bioinformatics researcher with the Genomic Surveillance of Antimicrobial Resistance project (GHRU-GSAR), building AMR genomics surveillance pipelines. I am a licensed pharmacist with the Pharmacists Council of Nigeria. The AIMS scholarship, funded by Google DeepMind, is designed to train African mathematicians who will solve African problems. Drug resistance is an African problem: the continent carries the highest burden of antimicrobial-resistant infections per capita, and most resistance prediction tools are trained and validated on Western sequence databases. My TOPOLOGIX work, and the AIMS training that would sharpen it, directly addresses that gap. I am applying to AIMS not as an alternative to my Potsdam degree but as a complement to it. The digital health master's gives me clinical data infrastructure and deployment skills. AIMS gives me the mathematical spine. Together they produce a researcher who can build, validate, and deploy quantitative biomedical tools from within Africa, for Africa. I ask for the scholarship to make that combination real. RESEARCH STATEMENT My research program asks one question: can advanced mathematics, applied rigorously, turn biological sequence data into clinically actionable predictions for African populations? The current answer, based on my work, is yes, but only if the mathematical methods are chosen and validated with discipline. I have spent the last two years building and testing computational models across pharmacology, neuroscience, and protein machine learning. The thread connecting them is a commitment to pre-registration, Bayesian calibration, and honest reporting of negative results. The centerpiece of my current work is TOPOLOGIX, a sequence-based predictor of drug-resistance mutations. The problem is concrete: structure-based tools like mCSM-lig require a resolved protein structure, which exists for only a fraction of clinically relevant mutations. On the Platinum benchmark of 553 mutations, structure-limited tools cover roughly 18 percent of cases. TOPOLOGIX uses ESM-2 protein language model delta-embeddings, which capture evolutionary and structural information from sequence alone, combined with Morgan/ECFP drug fingerprints and a Random Forest classifier. It achieves AUROC 0.804 plus or minus 0.025 on Platinum and 0.634 on SKEMPI 2.0, outperforming structure-based baselines while covering all mutations. The method is not a black box; the Random Forest provides feature importances, and I have published the full pipeline on GitHub for independent replication. TOPOLOGIX emerged from a falsified hypothesis. In 2024, I tested whether bipartite persistent homology, an algebraic topology method measuring opposition-distance geometry at protein-ligand interfaces, could predict hERG cardiotoxicity. The pre-registered, powered replication found it could not: AUROC 0.8426 versus 0.8782 for a plain descriptor baseline. I then applied the same topological constructs to drug-resistance prediction and found they carried almost no signal (AUROC 0.425 and 0.485 on the Platinum benchmark). These negative results, published as preprints and reported directly, ruled out interface geometry as the driver and motivated the sequence-representation approach that became TOPOLOGIX. This is the scientific method working as intended: hypothesis, test, falsification, pivot. My methodological toolkit extends beyond protein ML. In the CCT (Conjunctive Consolidation Threshold) model, I built a tripartite pharmacological framework for reward-memory encoding prevention in addiction, coupling dopaminergic reward prediction error, NMDAR-dependent long-term potentiation, and affective contrast in a system of ordinary differential equations solved with RK45. I calibrated the 14 free parameters using Bayesian MCMC (PyMC DEMetropolisZ) with priors elicited from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity of 13 to 22 percentage points across model versions. In neurocascade, I built a receptor-to-behavior simulation engine coupling pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readouts, with 62 of 62 tests passing and Bayesian calibration of the receptor and circuit layers. In ergofluids, I extended Koopman operator methods with a Mori-Zwanzig memory kernel to model drug transport through tumor tissue; the synthetic-data validation gates passed, and the first real-data gate did not meet its pre-registered criterion, which I reported directly rather than reframed. What AIMS offers me is the mathematical depth to push these methods further. The AIMS curriculum in algebra, analysis, and computational methods would strengthen my grasp of the representation theory underlying protein language models, the stochastic processes underlying Bayesian sampling, and the dynamical systems theory underlying my circuit models. The Google DeepMind scholarship specifically supports students who will apply mathematical training to real-world challenges. My record shows I already do that. AIMS would make me better at it, and I would bring that capacity back to Nigeria, where drug resistance, mental health infrastructure, and clinical data systems are urgent, under-resourced problems. ESSAY: COMMITMENT TO AFRICAN SCIENTIFIC DEVELOPMENT My commitment to African scientific development is documented in my work history, not just my intentions. As a bioinformatics researcher with the Genomic Surveillance of Antimicrobial Resistance project (GHRU-GSAR), I built AMR genomics surveillance pipelines that track resistance patterns in Nigerian clinical isolates. As a clinical pharmacist at Ramset Pharmacy and national product manager at Synthcare, I worked directly with the drug supply chain and clinical workflows that determine whether patients receive effective antibiotics. I am a licensed pharmacist with the Pharmacists Council of Nigeria, and I completed my B.Pharm at the University of Ibadan, one of the continent's leading research universities. The scientific problem I want to solve for Africa is drug resistance. The World Health Organization identifies antimicrobial resistance as one of the top ten global public health threats, and sub-Saharan Africa carries a disproportionate share of the burden. Resistance prediction tools are trained on databases like Platinum and SKEMPI 2.0, which are dominated by Western sequence data. A model trained on those data may not generalize to African clinical isolates, which have distinct resistance mechanisms and evolutionary pressures. My TOPOLOGIX project, which predicts resistance mutations from sequence alone, is designed to be retrainable on African data. The method does not require resolved protein structures, which are rarely available for African isolates, only sequence data, which are increasingly cheap to generate. AIMS is the right institution to support this work because it trains African mathematicians to solve African problems. The Google DeepMind scholarship adds a machine learning and AI dimension that directly matches my methods. My plan after AIMS is to return to Nigeria and build a computational biology research group focused on AMR prediction, drug repurposing, and clinical decision support. I have the research record, the technical skills, and the local knowledge to do this. What I need is the mathematical training to do it at the highest level. AIMS provides that training, and I will use it to build tools that serve African patients and African researchers. EDITOR NOTES - Research line chosen: TOPOLOGIX, the sequence-based drug-resistance predictor. This is the best fit for AIMS because it is the most mathematically grounded of my current projects (representation learning, classifier calibration, uncertainty quantification), it directly addresses African scientific development through AMR, and it is the current, active line, not a superseded or falsified one. The hERG topology study and the interface-topology-for-resistance study are mentioned only as falsified hypotheses that motivated TOPOLOGIX, which is honest and aligns with AIMS's emphasis on rigorous problem-solving. - Eligibility risk: The AIMS scholarship is primarily for graduate study at an AIMS center. The applicant is already enrolled in an M.Sc. at HPI/Potsdam. The motivation letter frames AIMS as a complement, not an alternative, but the applicant must verify whether AIMS allows concurrent enrollment or requires full-time attendance at an AIMS center. If full-time attendance is required, the Potsdam enrollment may need to be deferred or paused. This must be checked against the official AIMS and Google DeepMind scholarship terms before submission. - Facts to verify: The URL provided is a third-party aggregator (theeducationstory.com), not the official AIMS or Google DeepMind scholarship page. The applicant must confirm the official application portal, deadline, and required documents. The specific AIMS center (e.g., AIMS South Africa, AIMS Senegal, AIMS Rwanda) must be identified, as the application process and program structure vary by center. - Gaps to fill: The motivation letter and research statement do not name a specific AIMS center or faculty member. The applicant should research the faculty at their preferred AIMS center and reference one or two whose work aligns with TOPOLOGIX or computational biology. The essay on African scientific development should include one concrete anecdote from the GHRU-GSAR work, naming a specific pathogen, dataset, or surveillance finding, to ground the commitment in a specific fact. - Tone and formatting check: The draft uses first person, concrete numbers, and no banned phrases. The motivation letter opens with the AIMS model, not with "I." The research statement opens with the research question. The essay opens with the work history. All sections are within the specified word limits. The checklist below lists the documents required for a standard AIMS scholarship application; the applicant must confirm against the official portal. CHECKLIST - [ ] Confirm official AIMS and Google DeepMind scholarship 2026 application portal and deadline - [ ] Verify eligibility for concurrent enrollment with M.Sc. Digital Health at HPI/Potsdam - [ ] Identify preferred AIMS center (South Africa, Senegal, Rwanda, or other) and confirm program dates - [ ] Research AIMS faculty and add one or two named references to the motivation letter - [ ] Add one concrete GHRU-GSAR anecdote to the essay on African scientific development - [ ] Obtain official transcripts from University of Ibadan (B.Pharm) and HPI/Potsdam (current enrollment) - [ ] Obtain two academic or professional reference letters, one preferably from a mathematics or computational supervisor - [ ] Prepare CV in AIMS format, including ORCID, GitHub, and zyco.org links - [ ] Prepare PDF copies of TOPOLOGIX preprints and the hERG topology replication preprint - [ ] Prepare a one-page project proposal for TOPOLOGIX extension to African AMR data - [ ] Submit application through the official AIMS portal before the deadline - [ ] Prepare for possible interview or additional assessment, including a mathematical problem-solving component