Draft — Mozilla Fellowship Programme

MOTIVATION LETTER

The Mozilla Fellowship programme funds people who build technology that puts people first. I am an independent computational researcher based in Nigeria, and I build open, transparent tools for drug discovery and neuroscience modeling that are designed to be used by researchers who cannot afford enterprise software or high-performance computing clusters. My work sits at the intersection of addiction neuroscience, protein machine learning, and dynamical systems methods, and every tool I have released is publicly licensed, fully documented, and reproducible from source.

My primary research output is the Conjunctive Consolidation Threshold model, a tripartite pharmacological framework for preventing reward-memory encoding in addiction. The model couples dopaminergic reward prediction error, NMDA-receptor-dependent long-term potentiation, and affective contrast into a single system of ordinary differential equations. I calibrated the model using Bayesian Markov chain Monte Carlo methods with 14 free parameters and literature-elicited priors drawn from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity ranging from 13 to 22 percentage points across model versions. The three sole-authored preprints are archived on OSF and Zenodo, and a co-authored paper is under review at Alcohol.

My second major line is TOPOLOGIX, a protein-language-model-based platform that predicts drug-resistance mutations from sequence alone. It uses ESM-2 delta-embeddings combined with Morgan and ECFP drug fingerprints and a Random Forest classifier. On the Platinum benchmark of 553 mutations, TOPOLOGIX achieves an AUROC of 0.804 plus or minus 0.025. On SKEMPI 2.0, it achieves 0.634. These numbers beat structure-based baselines such as mCSM-lig, which scores roughly 0.70, while covering 100 percent of mutations compared to roughly 18 percent for structure-limited tools. The code is on GitHub under an open license.

I also maintain a self-hosted infrastructure stack that includes four independent DuckDB-based ingest-to-analyze pipelines spanning life sciences, technology and security, and social science domains. I serve local large language models using llama.cpp with on-demand model swapping, and I run production systems on Linux VPS with systemd, Caddy TLS, continuous integration and deployment, and automated backup and disaster recovery. This infrastructure is the operational backbone that lets me do independent research without institutional compute support.

The Mozilla Fellowship is the right programme for this work because Mozilla explicitly funds open infrastructure and people-centered technology. My tools are openly licensed, my data pipelines are reproducible, and my research questions are chosen to serve populations that commercial pharmaceutical research ignores: African communities facing addiction and antimicrobial resistance. I am enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute and the University of Potsdam starting winter semester 2026-2027, and I hold a B.Pharm from the University of Ibadan with a German-equivalent grade of 1.9. I am 29 years old, Nigerian, and eligible for early-career and LMIC-track programmes.

SHORT ESSAY: PROJECT PLAN

Over a 12-month Mozilla Fellowship, I will complete and release two openly licensed computational platforms and produce one peer-reviewed publication.

First, I will finalize TOPOLOGIX as a production-ready web tool. The current command-line Python implementation achieves an AUROC of 0.804 on the Platinum benchmark. I will build a browser-based interface that accepts a protein sequence and a drug SMILES string and returns a resistance prediction within seconds. The interface will be built with JavaScript and Node.js, backed by a Supabase and PostgreSQL database for storing user-submitted sequences and predictions. The entire stack will be deployed on my existing Linux VPS infrastructure with Caddy TLS and automated backups. The target is a public beta by month six.

Second, I will extend the neurocascade simulation engine to include a fourth receptor system: the serotonin 5-HT2A receptor, which is directly relevant to psychedelic-assisted therapy for addiction. The current engine couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readout, with three calibrated systems: mu-opioid, D2 dopamine, and GABA-A. All 62 of 62 unit tests pass. Adding 5-HT2A requires literature calibration of binding affinities, circuit-level parameters, and behavioral readout mapping. I will release the extended engine as a Python package on PyPI with full documentation and a tutorial notebook.

Third, I will submit a manuscript describing the CCT model to a peer-reviewed journal. The manuscript will report the full Bayesian calibration procedure, the confirmation of all five pre-registered hypotheses, and the posterior super-additivity results. The target journal is PLOS Computational Biology.

All code, data, and documentation will be released under permissive open-source licenses. The publication will be open access.

SHORT ESSAY: VALUES AND IMPACT

Mozilla's values of care, openness, and innovation in technology align directly with how I conduct research. Care means building tools that actually work for the people who need them. Openness means releasing every line of code, every data point, and every calibration prior so that anyone can verify, reuse, or challenge the results. Innovation means asking questions that the field has not asked, and reporting answers honestly even when they are negative.

My hERG cardiotoxicity study is an example. I tested whether bipartite persistent homology could predict hERG cardiotoxicity from protein-ligand interface geometry. This was a pre-registered, powered replication. The result was that topological features do not beat a plain descriptor baseline: AUROC 0.8426 versus 0.8782. The published literature had never actually run that comparison. I reported the negative result directly rather than reframing it. That is openness.

My interface-topology-for-resistance study is another example. I applied the same topological constructs to drug-resistance prediction and found they carry almost no signal: AUROC 0.425 and 0.485 on the Platinum benchmark. I reported that result and moved to a sequence-representation approach, which became TOPOLOGIX. That is innovation driven by honest failure.

The impact I aim for is specific. African researchers working on antimicrobial resistance currently rely on tools that cover only 18 percent of possible mutations because those tools require protein structures. TOPOLOGIX covers 100 percent of mutations from sequence alone. A researcher in Lagos or Nairobi with a laptop and an internet connection can use it. That is care.

I have already demonstrated leadership through speaking and collaboration. My endorsers include Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. I have presented my work at the University of Ibadan and through online preprint archives. I am building a network of independent researchers across Africa who share code and data through my self-hosted infrastructure.

CHECKLIST

- [ ] Motivation letter, 300-500 words, written in first person as Eniola Ayodele Olutogun
- [ ] Short essay on project plan, 200-350 words, describing 12-month deliverables
- [ ] Short essay on values and impact, 200-350 words, referencing Mozilla's criteria by name
- [ ] CV or resume, formatted for Mozilla Fellowship application
- [ ] Two letters of recommendation, one from a research collaborator and one from a professional reference
- [ ] Links to ORCID profile (0009-0001-9272-6735), GitHub (github.com/AmunRaPtah), and personal site (zyco.org)
- [ ] Links to preprints on OSF and Zenodo for CCT model
- [ ] Link to TOPOLOGIX repository on GitHub
- [ ] Proof of enrollment in M.Sc. Digital Health at Hasso Plattner Institute / University of Potsdam
- [ ] Proof of B.Pharm degree from University of Ibadan
- [ ] Nigerian passport or national ID for eligibility verification

EDITOR NOTES

- Eligibility risk: The Mozilla Fellowship typically requires full-time commitment. Eniola is enrolled in an M.Sc. programme starting winter 2026-2027 and holds a full-time position as National Product Manager at Synthcare. Verify whether the fellowship allows concurrent employment or study, or whether a leave of absence is required.
- Fact to verify: The Mozilla Fellowship application URL points to applynow.co.zw, which is a Zimbabwe-based job board. Confirm that this is the correct application portal and not a third-party aggregator. Check the official Mozilla Foundation website for the current fellowship cycle.
- Gap: The profile does not include any information about Eniola's experience with public speaking, conference presentations, or community organizing. Mozilla values leadership through speaking and writing. Eniola should prepare a list of talks, workshops, or community events he has led or participated in, even if informal.
- Gap: The profile does not specify whether Eniola has any prior experience with Mozilla programmes, events, or communities. If he has attended MozFest, contributed to Mozilla open-source projects, or participated in Mozilla study groups, that information should be added.
- Verification needed: The profile lists endorsements from four senior researchers. Confirm that these individuals are willing to write letters of recommendation or at minimum confirm the endorsement relationship. Mozilla may contact them.

MOTIVATION LETTER

The Mozilla Fellowship programme funds people who build technology that puts people first. I am an independent computational researcher based in Nigeria, and I build open, transparent tools for drug discovery and neuroscience modeling that are designed to be used by researchers who cannot afford enterprise software or high-performance computing clusters. My work sits at the intersection of addiction neuroscience, protein machine learning, and dynamical systems methods, and every tool I have released is publicly licensed, fully documented, and reproducible from source.

My primary research output is the Conjunctive Consolidation Threshold model, a tripartite pharmacological framework for preventing reward-memory encoding in addiction. The model couples dopaminergic reward prediction error, NMDA-receptor-dependent long-term potentiation, and affective contrast into a single system of ordinary differential equations. I calibrated the model using Bayesian Markov chain Monte Carlo methods with 14 free parameters and literature-elicited priors drawn from a systematic screen of 1,847 records. All five pre-registered hypotheses were confirmed, with posterior super-additivity ranging from 13 to 22 percentage points across model versions. The three sole-authored preprints are archived on OSF and Zenodo, and a co-authored paper is under review at Alcohol.

My second major line is TOPOLOGIX, a protein-language-model-based platform that predicts drug-resistance mutations from sequence alone. It uses ESM-2 delta-embeddings combined with Morgan and ECFP drug fingerprints and a Random Forest classifier. On the Platinum benchmark of 553 mutations, TOPOLOGIX achieves an AUROC of 0.804 plus or minus 0.025. On SKEMPI 2.0, it achieves 0.634. These numbers beat structure-based baselines such as mCSM-lig, which scores roughly 0.70, while covering 100 percent of mutations compared to roughly 18 percent for structure-limited tools. The code is on GitHub under an open license.

I also maintain a self-hosted infrastructure stack that includes four independent DuckDB-based ingest-to-analyze pipelines spanning life sciences, technology and security, and social science domains. I serve local large language models using llama.cpp with on-demand model swapping, and I run production systems on Linux VPS with systemd, Caddy TLS, continuous integration and deployment, and automated backup and disaster recovery. This infrastructure is the operational backbone that lets me do independent research without institutional compute support.

The Mozilla Fellowship is the right programme for this work because Mozilla explicitly funds open infrastructure and people-centered technology. My tools are openly licensed, my data pipelines are reproducible, and my research questions are chosen to serve populations that commercial pharmaceutical research ignores: African communities facing addiction and antimicrobial resistance. I am enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute and the University of Potsdam starting winter semester 2026-2027, and I hold a B.Pharm from the University of Ibadan with a German-equivalent grade of 1.9. I am 29 years old, Nigerian, and eligible for early-career and LMIC-track programmes.

SHORT ESSAY: PROJECT PLAN

Over a 12-month Mozilla Fellowship, I will complete and release two openly licensed computational platforms and produce one peer-reviewed publication.

First, I will finalize TOPOLOGIX as a production-ready web tool. The current command-line Python implementation achieves an AUROC of 0.804 on the Platinum benchmark. I will build a browser-based interface that accepts a protein sequence and a drug SMILES string and returns a resistance prediction within seconds. The interface will be built with JavaScript and Node.js, backed by a Supabase and PostgreSQL database for storing user-submitted sequences and predictions. The entire stack will be deployed on my existing Linux VPS infrastructure with Caddy TLS and automated backups. The target is a public beta by month six.

Second, I will extend the neurocascade simulation engine to include a fourth receptor system: the serotonin 5-HT2A receptor, which is directly relevant to psychedelic-assisted therapy for addiction. The current engine couples pharmacokinetics to receptor binding to Wilson-Cowan circuit dynamics to behavioral readout, with three calibrated systems: mu-opioid, D2 dopamine, and GABA-A. All 62 of 62 unit tests pass. Adding 5-HT2A requires literature calibration of binding affinities, circuit-level parameters, and behavioral readout mapping. I will release the extended engine as a Python package on PyPI with full documentation and a tutorial notebook.

Third, I will submit a manuscript describing the CCT model to a peer-reviewed journal. The manuscript will report the full Bayesian calibration procedure, the confirmation of all five pre-registered hypotheses, and the posterior super-additivity results. The target journal is PLOS Computational Biology.

All code, data, and documentation will be released under permissive open-source licenses. The publication will be open access.

SHORT ESSAY: VALUES AND IMPACT

Mozilla's values of care, openness, and innovation in technology align directly with how I conduct research. Care means building tools that actually work for the people who need them. Openness means releasing every line of code, every data point, and every calibration prior so that anyone can verify, reuse, or challenge the results. Innovation means asking questions that the field has not asked, and reporting answers honestly even when they are negative.

My hERG cardiotoxicity study is an example. I tested whether bipartite persistent homology could predict hERG cardiotoxicity from protein-ligand interface geometry. This was a pre-registered, powered replication. The result was that topological features do not beat a plain descriptor baseline: AUROC 0.8426 versus 0.8782. The published literature had never actually run that comparison. I reported the negative result directly rather than reframing it. That is openness.

My interface-topology-for-resistance study is another example. I applied the same topological constructs to drug-resistance prediction and found they carry almost no signal: AUROC 0.425 and 0.485 on the Platinum benchmark. I reported that result and moved to a sequence-representation approach, which became TOPOLOGIX. That is innovation driven by honest failure.

The impact I aim for is specific. African researchers working on antimicrobial resistance currently rely on tools that cover only 18 percent of possible mutations because those tools require protein structures. TOPOLOGIX covers 100 percent of mutations from sequence alone. A researcher in Lagos or Nairobi with a laptop and an internet connection can use it. That is care.

I have already demonstrated leadership through speaking and collaboration. My endorsers include Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. I have presented my work at the University of Ibadan and through online preprint archives. I am building a network of independent researchers across Africa who share code and data through my self-hosted infrastructure.

CHECKLIST

- [ ] Motivation letter, 300-500 words, written in first person as Eniola Ayodele Olutogun
- [ ] Short essay on project plan, 200-350 words, describing 12-month deliverables
- [ ] Short essay on values and impact, 200-350 words, referencing Mozilla's criteria by name
- [ ] CV or resume, formatted for Mozilla Fellowship application
- [ ] Two letters of recommendation, one from a research collaborator and one from a professional reference
- [ ] Links to ORCID profile (0009-0001-9272-6735), GitHub (github.com/AmunRaPtah), and personal site (zyco.org)
- [ ] Links to preprints on OSF and Zenodo for CCT model
- [ ] Link to TOPOLOGIX repository on GitHub
- [ ] Proof of enrollment in M.Sc. Digital Health at Hasso Plattner Institute / University of Potsdam
- [ ] Proof of B.Pharm degree from University of Ibadan
- [ ] Nigerian passport or national ID for eligibility verification

EDITOR NOTES

- Eligibility risk: The Mozilla Fellowship typically requires full-time commitment. Eniola is enrolled in an M.Sc. programme starting winter 2026-2027 and holds a full-time position as National Product Manager at Synthcare. Verify whether the fellowship allows concurrent employment or study, or whether a leave of absence is required.
- Fact to verify: The Mozilla Fellowship application URL points to applynow.co.zw, which is a Zimbabwe-based job board. Confirm that this is the correct application portal and not a third-party aggregator. Check the official Mozilla Foundation website for the current fellowship cycle.
- Gap: The profile does not include any information about Eniola's experience with public speaking, conference presentations, or community organizing. Mozilla values leadership through speaking and writing. Eniola should prepare a list of talks, workshops, or community events he has led or participated in, even if informal.
- Gap: The profile does not specify whether Eniola has any prior experience with Mozilla programmes, events, or communities. If he has attended MozFest, contributed to Mozilla open-source projects, or participated in Mozilla study groups, that information should be added.
- Verification needed: The profile lists endorsements from four senior researchers. Confirm that these individuals are willing to write letters of recommendation or at minimum confirm the endorsement relationship. Mozilla may contact them.

AI Draft — Mozilla Fellowship Programme