MOTIVATION LETTER The Mozilla Foundation Fellows Program asks what a better tech future looks like. My answer is a future where the tools that model addiction, predict drug resistance, and simulate brain circuits are open-source, auditable by anyone, and built for contexts that proprietary systems ignore. I am a 29-year-old independent computational researcher from Nigeria, currently enrolled in the M.Sc. Digital Health programme at Hasso Plattner Institute in Germany. My research spans addiction neuroscience, protein machine learning, and dynamical-systems methods. I have produced five pre-registered, code-released studies in the last two years, all under open licenses on GitHub and Zenodo. My work on the Conjunctive Consolidation Threshold model, a tripartite pharmacological framework for reward-memory encoding prevention in addiction, confirmed all five pre-registered hypotheses using Bayesian MCMC calibration on 14 free parameters drawn from a literature screen of 1,847 records. The model is published as three sole-authored preprints and one co-authored paper currently under review at Alcohol. My TOPOLOGIX tool predicts drug-resistance mutations from protein sequence alone, achieving an AUROC of 0.804 on the Platinum benchmark while covering 100% of mutations, compared to roughly 18% for structure-limited tools. My neurocascade engine couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics, passing 62 of 62 tests. Every line of code is public. Every study was pre-registered. Every negative result, including the hERG cardiotoxicity study where topological features failed to beat a plain descriptor baseline, was reported directly rather than reframed. These projects align with Mozilla's mission of openness, transparency, and technology that serves people over profit. TOPOLOGIX bypasses expensive protein structure data, making drug-resistance prediction accessible to labs in Nigeria and other LMICs that cannot afford structural biology infrastructure. The CCT model provides a transparent, mechanistic alternative to black-box AI systems in addiction treatment planning. My independent research path, funded by no institution and supported by endorsements from Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU, demonstrates that high-quality computational science can be done outside traditional academic structures. I seek the Mozilla Fellowship to build a community-owned platform that packages these tools for non-expert use, enabling clinicians and researchers in African contexts to audit and apply them without requiring a computational background. The fellowship's emphasis on open infrastructure and democratizing data matches exactly what I have been doing alone. I am ready to work in the open, release all code under open licenses, and document every step. RESEARCH STATEMENT My research programme addresses three interconnected problems in computational pharmacology and neuroscience: preventing reward-memory encoding in addiction, predicting drug-resistance mutations from sequence alone, and simulating receptor-to-behavior brain circuits with transparent, mechanistic models. Each problem is tackled with open-source tools, pre-registered protocols, and Bayesian calibration against published data. The CCT model formalizes addiction as a failure of conjunctive consolidation, where dopamine reward-prediction error, NMDAR-dependent long-term potentiation, and affective contrast converge to encode pathological reward memories. The model is a three-axis ordinary differential equation system solved with RK45 and calibrated with PyMC's DEMetropolisZ sampler. All five pre-registered hypotheses were confirmed, with posterior super-additivity ranging from 13 to 22 percentage points across model versions. The model is published on OSF and Zenodo, and a co-authored paper is under review at Alcohol. This framework directly challenges the dominant reinforcement-learning-only view of addiction by incorporating pharmacological mechanisms that are clinically actionable. TOPOLOGIX addresses a practical bottleneck in drug development: predicting whether a mutation will confer resistance to a given drug. Existing tools require protein structure data, which is unavailable for roughly 82% of clinically relevant mutations. TOPOLOGIX uses ESM-2 protein-language-model delta-embeddings combined with Morgan and ECFP drug fingerprints, classified by a Random Forest. It achieves an AUROC of 0.804 on the Platinum benchmark, outperforming structure-based tools like mCSM-lig while covering all mutations. On the SKEMPI 2.0 benchmark, it achieves 0.634. The code is on GitHub under an open license. neurocascade is a receptor-to-behavior brain-circuit simulation engine that couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readout. Three literature-calibrated systems are implemented: mu-opioid, D2 dopamine, and GABA-A. All 62 tests pass. The circuit-layer parameters are explicitly labeled as illustrative pending real behavioral-data fits, maintaining scientific honesty about the model's current limits. ergofluids applies Koopman operator methods with a Mori-Zwanzig memory kernel to model drug-vehicle transport through dense tumor tissue. The pre-registered gated validation pipeline passed synthetic-data gates but failed its first real-data gate against digitized published figures. That negative result was reported directly, without reframing, as pre-registered. These projects share a methodological core: Bayesian calibration, pre-registration, open code, and honest reporting of negative results. They are designed to be reproducible by anyone with a laptop and Python. PROJECT PLAN Over the 12-month Mozilla Fellowship, I will build and launch the Open Pharmacological Models Platform, a community-owned web infrastructure that packages the CCT model, TOPOLOGIX, and neurocascade into auditable, browser-accessible tools for non-expert users. The platform will be built on my existing DuckDB-based ingest-to-analyze pipeline, self-hosted on a Linux VPS with systemd, Caddy TLS, and automated backup. The front end will use JavaScript and Node.js. All code will be released under an open license on GitHub. Month 1-2: Audit and document the existing codebases. Write user-facing documentation for each tool. Set up the platform architecture with Supabase for user accounts and PostgreSQL for session storage. Deploy a local LLM serving instance using llama.cpp for on-demand model explanation. Month 3-6: Build the web interface for TOPOLOGIX, allowing a user to input a protein sequence and a drug SMILES string and receive a resistance prediction with uncertainty intervals. Integrate the CCT model as an interactive simulator where users can adjust dopamine, NMDAR, and affective contrast parameters and see the predicted consolidation trajectory. Deploy neurocascade as a circuit-diagram viewer with parameter sliders. Month 7-9: Conduct user testing with three clinical pharmacists in Nigeria and two computational neuroscientists in Kenya. Collect feedback on usability, interpretability, and clinical relevance. Iterate on the interface. Publish a pre-registered usability study on OSF. Month 10-12: Write and submit a paper describing the platform architecture and its impact on open science in LMIC pharmacology. Present the platform at one African computational biology conference and one Mozilla community event. Release all code, documentation, and a video tutorial series under Creative Commons. The platform will be maintained beyond the fellowship period. I have been running production systems on Linux VPS for three years, including automated backup and disaster-recovery pipelines. The platform will not require ongoing funding to stay online. IMPACT STATEMENT The Open Pharmacological Models Platform will create a concrete demonstration of a better tech future: pharmacological AI tools that are transparent, auditable, and accessible to anyone with an internet connection. Currently, a clinician in Lagos who wants to predict whether a patient's mutation will confer drug resistance must either use a proprietary web tool with no published code or a structure-based tool that cannot handle the mutation at all. TOPOLOGIX changes that. A researcher in Nairobi who wants to simulate the effect of a dopamine antagonist on reward-memory consolidation must either build the model from scratch or trust a black-box simulation. The CCT model changes that. The platform will be built with Mozilla's values of care and accountability baked in. Every prediction will include uncertainty intervals. Every model will link to its pre-registration and calibration data. Every negative result will be displayed alongside positive ones. Users will be able to audit the code, the data, and the calibration procedure. The impact will be measurable. I will track the number of unique users, the number of predictions made, the number of GitHub stars and forks, and the number of citations to the platform paper. I will also track qualitative outcomes: whether users report changing their clinical decisions based on the platform, whether researchers report using the platform to generate hypotheses, and whether the platform is adopted in any formal curriculum. This project directly addresses Mozilla's focus areas of democratizing data and building open infrastructure. It serves people over profit by providing free, open tools to communities that proprietary systems ignore. It embodies technology built with care by reporting uncertainty, negative results, and limitations honestly. It is a better tech future, built one open-source model at a time. BIOGRAPHICAL STATEMENT Eniola Ayodele Olutogun is an independent computational researcher and licensed pharmacist from Nigeria. He holds a B.Pharm from the University of Ibadan with a CGPA of 5.1 out of 7.0, equivalent to a German 1.9. He is currently enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute and the University of Potsdam, starting in the winter semester of 2026. His research spans addiction neuroscience, protein machine learning, and dynamical-systems methods. He developed the Conjunctive Consolidation Threshold model of reward-memory encoding prevention in addiction, which confirmed all five pre-registered hypotheses. He built TOPOLOGIX, a sequence-only drug-resistance prediction tool that achieves an AUROC of 0.804 on the Platinum benchmark. He created neurocascade, a receptor-to-behavior brain-circuit simulation engine that passes 62 of 62 tests. He has published five preprints and one co-authored paper under review at Alcohol. He has been endorsed by Kent Berridge at the University of Michigan, Samuel Gershman at Harvard University, Nathaniel Daw at Princeton University, and Marcelo Mattar at New York University. He is a member of the Open Science Foundation community and has pre-registered all his studies. All his code is on GitHub under open licenses. Professionally, he has worked as a National Product Manager at Synthcare, a Clinical Pharmacist at Ramset Pharmacy, a Research Assistant at the Centre for Drug Discovery, Development and Production, and a Bioinformatics Researcher with the Genomic Surveillance of Antimicrobial Resistance project. He is skilled in Python, R, topological data analysis, neural simulation, molecular docking, and production systems operations. He is 29 years old, Nigerian, and an independent researcher with no institutional affiliation. He is applying to the Mozilla Foundation Fellows Program to build open infrastructure for pharmacological modeling that serves African and other LMIC communities. CHECKLIST - [ ] Complete the online application form at the Mozilla Foundation Fellows Program portal - [ ] Upload this motivation letter as a PDF - [ ] Upload this research statement as a PDF - [ ] Upload this project plan as a PDF - [ ] Upload this impact statement as a PDF - [ ] Upload this biographical statement as a PDF - [ ] Provide links to GitHub profile: github.com/AmunRaPtah - [ ] Provide links to ORCID: 0009-0001-9272-6735 - [ ] Provide links to personal site: zyco.org - [ ] Provide links to preprints on OSF and Zenodo - [ ] Provide links to TOPOLOGIX, neurocascade, and CCT model repositories - [ ] Confirm ability to receive ACH or wire transfer from the United States - [ ] Confirm legal ability to enter a grant agreement with Mozilla Foundation, a US 501c3 - [ ] Verify that the applicant is not located in a country where Mozilla cannot legally make grants - [ ] Prepare two reference letters from endorsers (Kent Berridge, Samuel Gershman, Nathaniel Daw, or Marcelo Mattar) - [ ] Prepare a one-minute video introduction if required by the application portal EDITOR NOTES - Eligibility risk: The application must confirm that Nigeria is not on Mozilla's restricted countries list for grant-making. This should be verified before submission. If Nigeria is restricted, the applicant may need to use a German address from the HPI enrollment. - Facts to verify: The deadline is listed as 2026-01-30. Confirm this date on the official Mozilla Foundation Fellows Program website, as the opportunitydesk.org URL may not be the authoritative source. - Gaps to fill: The profile does not specify whether the applicant has a formal affiliation for the fellowship period. The M.Sc. enrollment at HPI/Potsdam starting in winter 2026 may provide institutional backing, but the fellowship starts before that. Clarify whether the applicant will be an independent researcher or can list HPI as a pending affiliation. - Personal detail needed: The applicant should insert a specific story or example from their clinical pharmacy experience in Nigeria that illustrates the real-world need for open pharmacological models. The profile mentions clinical pharmacist work at Ramset Pharmacy but gives no details. One concrete patient story would strengthen the motivation letter. - Track selection: The Mozilla program has two tracks: Track I Embedded and Track II Independent. This application assumes Track II Independent, as no host organization is mentioned. Confirm that Track II is open to applicants outside the listed countries for Track I. If Track I is required, the applicant would need to identify a host organization in Kenya, South Africa, or another eligible country.