Draft — Energy, Water, and Resource Engineering

MOTIVATION LETTER

The Energy, Water, and Resource Engineering programme at NSF funds fundamental engineering research that protects public health through environmental systems. My independent research in computational pharmacology has produced a validated mathematical framework for predicting neurochemical disruption in human reward circuits. That framework, the Conjunctive Consolidation Threshold model, is directly transferable to detecting neurotoxic compounds in water and soil systems. I am applying to this programme because my work bridges computational neuroscience and environmental engineering in a way that NSF EWRE explicitly supports.

I am Eniola Ayodele Olutogun, an independent researcher based in Lagos, Nigeria. My CCT model, published as three sole-authored preprints on OSF and Zenodo, uses ordinary differential equations and Bayesian Markov Chain Monte Carlo methods to predict how small molecules alter dopamine-mediated memory encoding. The model achieved an 85.8 percent reduction in encoding probability under combined pharmacological intervention, with super-additivity of 12.8 percentage points beyond individual drug effects. All five pre-registered hypotheses were confirmed. This mathematical architecture, validated against known pharmacology, can be repurposed to predict environmental neurotoxins.

My technical platforms demonstrate the engineering translation. TOPOLOGIX applies topological data analysis, persistent homology, and bipartite simplicial complexes to map drug-protein interactions. I built a minimum viable product for hERG cardiotoxicity screening using this method. IMPRINT screens compounds for addiction liability. These tools use the same computational stack: Python with scipy, numpy, PyMC, and Gudhi for TDA, running on HPC clusters with Nextflow and SLURM. The same pipeline can detect organophosphates, heavy metals, or endocrine disruptors in water samples by modeling their protein-binding signatures.

Nigeria faces severe environmental health challenges. Industrial discharge into the Lagos Lagoon, agricultural runoff containing neurotoxic pesticides, and informal e-waste processing expose millions to compounds that damage neural development and cognitive function. No systematic computational screening exists for these risks. My CCT framework and TDA tools can fill that gap. I have provisional patent protection on the core CCT architecture, filed Q3 2026, and endorsements from Kent Berridge at University of Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU.

I seek a U.S. academic collaborator, ideally at University of Michigan or Georgia Tech, to serve as principal investigator and submitting organization. My role would be co-investigator and computational lead, adapting my models to environmental contaminant detection. The EWRE programme supports exactly this kind of fundamental engineering research on AI-driven risk assessment for environmental health. I am not yet enrolled in a graduate programme, but I will apply for October 2026 entry at Medical University of Graz, Austria. This grant would fund the computational development and validation work between now and that enrollment.

RESEARCH STATEMENT

The Conjunctive Consolidation Threshold model is a tripartite pharmacological framework that predicts how combinations of drugs alter the probability of reward-memory encoding in the brain. I developed this model as an independent researcher in Lagos, Nigeria, between 2025 and 2026. The model combines three components: a dopamine-dependent gating function, a calcium-mediated plasticity threshold, and a protein synthesis consolidation gate. Each component is represented as a differential equation solved with Runge-Kutta 45 methods. The full system is calibrated using Bayesian MCMC sampling against published electrophysiology and behavioral data.

The model produces a single output: the probability that a given drug combination will prevent the encoding of a reward-associated memory. In my validation experiments, the model reduced encoding probability from 0.855 to 0.122, an 85.8 percent reduction. The combined effect of three drugs exceeded the sum of individual effects by 12.8 percentage points, demonstrating super-additivity. These results were pre-registered as hypotheses H1 through H5 and confirmed in full. A review article describing the model is under review at Neuroscience and Biobehavioral Reviews. A co-authored paper on alcohol pharmacology is under review at Alcohol (Elsevier).

For the EWRE programme, I propose to reframe this computational architecture as an environmental neurotoxin detection system. The same mathematical framework that predicts how drugs interact with dopamine receptors and calcium channels can predict how environmental contaminants interact with neural proteins. The key adaptation is replacing known drug pharmacodynamics with predicted binding affinities from molecular docking simulations. I have experience with AutoDock, RDKit, and ADMET/QSAR pipelines from my work at the Centre for Drug Discovery, Development and Production, where I performed NMDA receptor and insulin receptor docking studies.

The technical implementation uses TOPOLOGIX, my topological data analysis platform. TOPOLOGIX applies persistent homology to bipartite simplicial complexes representing drug-protein interaction networks. For environmental applications, the input would be a library of known and suspected water contaminants, their predicted protein targets from AlphaFold structures, and their binding affinities from molecular dynamics simulations using GROMACS. The output would be a risk score for each contaminant, indicating its probability of disrupting neural reward-memory encoding. This is a direct engineering application of the CCT model's mathematical core.

The validation strategy has three phases. Phase one: benchmark against 50 known neurotoxins with established human exposure data from the U.S. Environmental Protection Agency's ToxCast database. Phase two: blind prediction of neurotoxicity for 100 compounds not in the training set, with experimental validation using published in vitro assays. Phase three: field deployment in Lagos Lagoon, collecting water samples from 20 sites and comparing computational predictions against mass spectrometry analysis. This phased approach matches NSF EWRE's emphasis on fundamental engineering research with clear validation milestones.

The broader impact addresses a critical gap in environmental health monitoring for sub-Saharan Africa. Nigeria has no systematic computational screening for neurotoxic contaminants in water systems. The Lagos Lagoon receives untreated industrial effluent from over 10,000 factories. Agricultural runoff from cocoa, cassava, and palm oil plantations carries organophosphate pesticides. Informal e-waste recycling in Alaba International Market releases lead, cadmium, and brominated flame retardants into groundwater. My computational tools can prioritize the most dangerous compounds for targeted monitoring, reducing the cost and time of environmental surveillance.

I have provisional patent protection on the CCT core architecture, filed Q3 2026. I have endorsements from four leading computational neuroscientists: Kent Berridge at University of Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. Samuel Gershman provided my arXiv endorsement. I am seeking a U.S. academic collaborator at University of Michigan or Georgia Tech to serve as principal investigator. My role would be co-investigator and computational lead, responsible for model development, validation, and field deployment coordination. The EWRE programme's support for AI modeling in environmental health makes this a natural fit.

PROJECT NARRATIVE

Problem: Environmental neurotoxins in water systems are poorly characterized in sub-Saharan Africa. Existing monitoring relies on targeted chemical analysis, which misses unknown compounds and ignores synergistic effects. The Lagos Lagoon alone receives over 200,000 cubic meters of untreated industrial effluent daily. No computational tool exists to predict which compounds pose the greatest risk to neural development and cognitive function.

Approach: Adapt the Conjunctive Consolidation Threshold model, a validated computational pharmacology framework, to predict environmental neurotoxicity. The CCT model uses ordinary differential equations and Bayesian MCMC to calculate the probability that a compound disrupts reward-memory encoding. For environmental applications, replace known drug pharmacodynamics with predicted binding affinities from molecular docking. Use TOPOLOGIX, my topological data analysis platform, to map compound-protein interaction networks using persistent homology and bipartite simplicial complexes.

Methods: Phase one: compile a training set of 50 known neurotoxins from EPA ToxCast, with published human exposure data. For each compound, perform molecular docking against 20 neural protein targets using AutoDock Vina. Run 100-nanosecond molecular dynamics simulations in GROMACS for the top 10 binding poses. Extract binding free energies and residence times. Feed these parameters into the CCT model to predict encoding probability. Compare predictions against published in vitro neurotoxicity data. Phase two: blind test on 100 compounds not in the training set. Phase three: collect water samples from 20 sites in Lagos Lagoon, analyze with liquid chromatography-mass spectrometry, compare detected compounds against computational predictions.

Timeline: Months 1-6: compile training set, perform docking and dynamics simulations, calibrate CCT model for environmental compounds. Months 7-12: blind validation on 100 compounds, publish results. Months 13-18: field deployment in Lagos Lagoon, sample collection and analysis. Months 19-24: integrate computational and experimental results, publish final model, release open-source software.

Deliverables: An open-source computational pipeline for predicting environmental neurotoxicity from chemical structure. A validated risk score for 150 compounds. A field dataset from Lagos Lagoon with computational predictions and mass spectrometry confirmation. A peer-reviewed publication in an environmental engineering journal.

Budget: Computational resources for molecular dynamics simulations and HPC access, 15,000 USD. Field sampling equipment and mass spectrometry analysis, 25,000 USD. Travel for collaborator meetings and field work, 10,000 USD. Open-access publication fees, 3,000 USD. Total: 53,000 USD.

PERSONAL STATEMENT

I am a 29-year-old Nigerian pharmacist and independent computational researcher. I earned my Bachelor of Pharmacy from the University of Ibadan in 2021 with a CGPA of 5.1 out of 7.0, a German equivalent of 1.9. I am licensed by the Pharmacists Council of Nigeria. I have worked as a clinical pharmacist at Ramset Pharmacy, as National Product Manager at Synthcare, and as a research assistant at the Centre for Drug Discovery, Design and Development. I currently serve as a bioinformatics researcher with the Genomic Surveillance of Antimicrobial Resistance project, building surveillance pipelines for AMR genomics.

My research trajectory has been entirely self-directed. Between 2025 and 2026, working from Lagos without institutional affiliation, I developed the Conjunctive Consolidation Threshold model. I taught myself Bayesian statistics, ordinary differential equations, and topological data analysis. I wrote three sole-authored preprints, built two computational platforms, and filed a provisional patent. I secured endorsements from four leading computational neuroscientists at Michigan, Harvard, Princeton, and NYU. I did this without a supervisor, without a lab, and without research funding.

Nigeria shapes my research priorities. The country has 200 million people, fewer than 500 practicing psychiatrists, and no systematic environmental neurotoxin monitoring. My work addresses problems that matter here: addiction, environmental contamination, and the absence of computational tools for public health decision-making. The CCT model was motivated by the rising rate of substance use disorders among Nigerian youth. The environmental application was motivated by the Lagos Lagoon crisis, where industrial pollution has made fishing unsafe and drinking water unreliable.

I am applying to graduate programmes for October 2026 entry, with Medical University of Graz in Austria as my primary target. The EWRE grant would fund the computational development and field validation work between now and that enrollment. I am eligible for early-career, pre-PhD, and LMIC-track funding programmes. My long-term goal is to establish a computational environmental health laboratory in Lagos, training Nigerian scientists in AI-driven risk assessment and building tools that serve African communities.

CHECKLIST

- [ ] Identify a U.S. academic collaborator at University of Michigan or Georgia Tech to serve as PI and submitting organization
- [ ] Confirm collaborator's willingness to submit to NSF EWRE programme
- [ ] Verify collaborator's eligibility as NSF PI (U.S. institution, faculty or research scientist status)
- [ ] Draft collaborator letter of support describing their role and institutional commitment
- [ ] Prepare biosketch for Eniola Ayodele Olutogun (NSF format, 2-page limit)
- [ ] Prepare biosketch for collaborator
- [ ] Write project summary (1-page, NSF format)
- [ ] Write project description (15-page maximum, NSF format)
- [ ] Write budget justification with detailed cost breakdown
- [ ] Compile references cited (no page limit)
- [ ] Gather supporting documents: ORCID profile, GitHub repositories, preprint links, patent filing receipt
- [ ] Obtain letters of endorsement from Kent Berridge, Samuel Gershman, Nathaniel Daw, Marcelo Mattar
- [ ] Verify NSF EWRE deadline on programme website
- [ ] Confirm whether NSF requires U.S. citizenship or permanent residency for co-investigators
- [ ] Check if NSF allows foreign organizations as subawardees or if collaborator must manage all funds

EDITOR NOTES

- Eligibility risk: NSF EWRE typically requires the PI to be at a U.S. institution. Eniola cannot be PI. The collaborator must be confirmed before submission. If no collaborator is secured, this application cannot proceed.
- Fact to verify: Confirm that NSF allows foreign nationals as co-investigators on standard research grants. Some NSF programmes restrict co-PI eligibility to U.S. citizens or permanent residents. Check the NSF Proposal and Award Policies and Procedures Guide.
- Gap to fill: The profile does not specify which U.S. collaborator has been contacted or agreed to participate. This must be resolved before writing the full proposal. University of Michigan and Georgia Tech are suggested, but no confirmation exists.
- Budget detail needed: The 53,000 USD estimate is preliminary. NSF requires a detailed budget with specific cost categories, fringe benefits, indirect costs, and institutional facilities and administrative costs. The collaborator's institution will set the indirect cost rate.
- Missing personal detail: The profile does not include Eniola's date of birth, which may be required for NSF biographical sketches. Also missing: any prior grant funding, teaching experience, or awards that could strengthen the personal statement.

MOTIVATION LETTER

The Energy, Water, and Resource Engineering programme at NSF funds fundamental engineering research that protects public health through environmental systems. My independent research in computational pharmacology has produced a validated mathematical framework for predicting neurochemical disruption in human reward circuits. That framework, the Conjunctive Consolidation Threshold model, is directly transferable to detecting neurotoxic compounds in water and soil systems. I am applying to this programme because my work bridges computational neuroscience and environmental engineering in a way that NSF EWRE explicitly supports.

I am Eniola Ayodele Olutogun, an independent researcher based in Lagos, Nigeria. My CCT model, published as three sole-authored preprints on OSF and Zenodo, uses ordinary differential equations and Bayesian Markov Chain Monte Carlo methods to predict how small molecules alter dopamine-mediated memory encoding. The model achieved an 85.8 percent reduction in encoding probability under combined pharmacological intervention, with super-additivity of 12.8 percentage points beyond individual drug effects. All five pre-registered hypotheses were confirmed. This mathematical architecture, validated against known pharmacology, can be repurposed to predict environmental neurotoxins.

My technical platforms demonstrate the engineering translation. TOPOLOGIX applies topological data analysis, persistent homology, and bipartite simplicial complexes to map drug-protein interactions. I built a minimum viable product for hERG cardiotoxicity screening using this method. IMPRINT screens compounds for addiction liability. These tools use the same computational stack: Python with scipy, numpy, PyMC, and Gudhi for TDA, running on HPC clusters with Nextflow and SLURM. The same pipeline can detect organophosphates, heavy metals, or endocrine disruptors in water samples by modeling their protein-binding signatures.

Nigeria faces severe environmental health challenges. Industrial discharge into the Lagos Lagoon, agricultural runoff containing neurotoxic pesticides, and informal e-waste processing expose millions to compounds that damage neural development and cognitive function. No systematic computational screening exists for these risks. My CCT framework and TDA tools can fill that gap. I have provisional patent protection on the core CCT architecture, filed Q3 2026, and endorsements from Kent Berridge at University of Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU.

I seek a U.S. academic collaborator, ideally at University of Michigan or Georgia Tech, to serve as principal investigator and submitting organization. My role would be co-investigator and computational lead, adapting my models to environmental contaminant detection. The EWRE programme supports exactly this kind of fundamental engineering research on AI-driven risk assessment for environmental health. I am not yet enrolled in a graduate programme, but I will apply for October 2026 entry at Medical University of Graz, Austria. This grant would fund the computational development and validation work between now and that enrollment.

RESEARCH STATEMENT

The Conjunctive Consolidation Threshold model is a tripartite pharmacological framework that predicts how combinations of drugs alter the probability of reward-memory encoding in the brain. I developed this model as an independent researcher in Lagos, Nigeria, between 2025 and 2026. The model combines three components: a dopamine-dependent gating function, a calcium-mediated plasticity threshold, and a protein synthesis consolidation gate. Each component is represented as a differential equation solved with Runge-Kutta 45 methods. The full system is calibrated using Bayesian MCMC sampling against published electrophysiology and behavioral data.

The model produces a single output: the probability that a given drug combination will prevent the encoding of a reward-associated memory. In my validation experiments, the model reduced encoding probability from 0.855 to 0.122, an 85.8 percent reduction. The combined effect of three drugs exceeded the sum of individual effects by 12.8 percentage points, demonstrating super-additivity. These results were pre-registered as hypotheses H1 through H5 and confirmed in full. A review article describing the model is under review at Neuroscience and Biobehavioral Reviews. A co-authored paper on alcohol pharmacology is under review at Alcohol (Elsevier).

For the EWRE programme, I propose to reframe this computational architecture as an environmental neurotoxin detection system. The same mathematical framework that predicts how drugs interact with dopamine receptors and calcium channels can predict how environmental contaminants interact with neural proteins. The key adaptation is replacing known drug pharmacodynamics with predicted binding affinities from molecular docking simulations. I have experience with AutoDock, RDKit, and ADMET/QSAR pipelines from my work at the Centre for Drug Discovery, Development and Production, where I performed NMDA receptor and insulin receptor docking studies.

The technical implementation uses TOPOLOGIX, my topological data analysis platform. TOPOLOGIX applies persistent homology to bipartite simplicial complexes representing drug-protein interaction networks. For environmental applications, the input would be a library of known and suspected water contaminants, their predicted protein targets from AlphaFold structures, and their binding affinities from molecular dynamics simulations using GROMACS. The output would be a risk score for each contaminant, indicating its probability of disrupting neural reward-memory encoding. This is a direct engineering application of the CCT model's mathematical core.

The validation strategy has three phases. Phase one: benchmark against 50 known neurotoxins with established human exposure data from the U.S. Environmental Protection Agency's ToxCast database. Phase two: blind prediction of neurotoxicity for 100 compounds not in the training set, with experimental validation using published in vitro assays. Phase three: field deployment in Lagos Lagoon, collecting water samples from 20 sites and comparing computational predictions against mass spectrometry analysis. This phased approach matches NSF EWRE's emphasis on fundamental engineering research with clear validation milestones.

The broader impact addresses a critical gap in environmental health monitoring for sub-Saharan Africa. Nigeria has no systematic computational screening for neurotoxic contaminants in water systems. The Lagos Lagoon receives untreated industrial effluent from over 10,000 factories. Agricultural runoff from cocoa, cassava, and palm oil plantations carries organophosphate pesticides. Informal e-waste recycling in Alaba International Market releases lead, cadmium, and brominated flame retardants into groundwater. My computational tools can prioritize the most dangerous compounds for targeted monitoring, reducing the cost and time of environmental surveillance.

I have provisional patent protection on the CCT core architecture, filed Q3 2026. I have endorsements from four leading computational neuroscientists: Kent Berridge at University of Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. Samuel Gershman provided my arXiv endorsement. I am seeking a U.S. academic collaborator at University of Michigan or Georgia Tech to serve as principal investigator. My role would be co-investigator and computational lead, responsible for model development, validation, and field deployment coordination. The EWRE programme's support for AI modeling in environmental health makes this a natural fit.

PROJECT NARRATIVE

Problem: Environmental neurotoxins in water systems are poorly characterized in sub-Saharan Africa. Existing monitoring relies on targeted chemical analysis, which misses unknown compounds and ignores synergistic effects. The Lagos Lagoon alone receives over 200,000 cubic meters of untreated industrial effluent daily. No computational tool exists to predict which compounds pose the greatest risk to neural development and cognitive function.

Approach: Adapt the Conjunctive Consolidation Threshold model, a validated computational pharmacology framework, to predict environmental neurotoxicity. The CCT model uses ordinary differential equations and Bayesian MCMC to calculate the probability that a compound disrupts reward-memory encoding. For environmental applications, replace known drug pharmacodynamics with predicted binding affinities from molecular docking. Use TOPOLOGIX, my topological data analysis platform, to map compound-protein interaction networks using persistent homology and bipartite simplicial complexes.

Methods: Phase one: compile a training set of 50 known neurotoxins from EPA ToxCast, with published human exposure data. For each compound, perform molecular docking against 20 neural protein targets using AutoDock Vina. Run 100-nanosecond molecular dynamics simulations in GROMACS for the top 10 binding poses. Extract binding free energies and residence times. Feed these parameters into the CCT model to predict encoding probability. Compare predictions against published in vitro neurotoxicity data. Phase two: blind test on 100 compounds not in the training set. Phase three: collect water samples from 20 sites in Lagos Lagoon, analyze with liquid chromatography-mass spectrometry, compare detected compounds against computational predictions.

Timeline: Months 1-6: compile training set, perform docking and dynamics simulations, calibrate CCT model for environmental compounds. Months 7-12: blind validation on 100 compounds, publish results. Months 13-18: field deployment in Lagos Lagoon, sample collection and analysis. Months 19-24: integrate computational and experimental results, publish final model, release open-source software.

Deliverables: An open-source computational pipeline for predicting environmental neurotoxicity from chemical structure. A validated risk score for 150 compounds. A field dataset from Lagos Lagoon with computational predictions and mass spectrometry confirmation. A peer-reviewed publication in an environmental engineering journal.

Budget: Computational resources for molecular dynamics simulations and HPC access, 15,000 USD. Field sampling equipment and mass spectrometry analysis, 25,000 USD. Travel for collaborator meetings and field work, 10,000 USD. Open-access publication fees, 3,000 USD. Total: 53,000 USD.

PERSONAL STATEMENT

I am a 29-year-old Nigerian pharmacist and independent computational researcher. I earned my Bachelor of Pharmacy from the University of Ibadan in 2021 with a CGPA of 5.1 out of 7.0, a German equivalent of 1.9. I am licensed by the Pharmacists Council of Nigeria. I have worked as a clinical pharmacist at Ramset Pharmacy, as National Product Manager at Synthcare, and as a research assistant at the Centre for Drug Discovery, Design and Development. I currently serve as a bioinformatics researcher with the Genomic Surveillance of Antimicrobial Resistance project, building surveillance pipelines for AMR genomics.

My research trajectory has been entirely self-directed. Between 2025 and 2026, working from Lagos without institutional affiliation, I developed the Conjunctive Consolidation Threshold model. I taught myself Bayesian statistics, ordinary differential equations, and topological data analysis. I wrote three sole-authored preprints, built two computational platforms, and filed a provisional patent. I secured endorsements from four leading computational neuroscientists at Michigan, Harvard, Princeton, and NYU. I did this without a supervisor, without a lab, and without research funding.

Nigeria shapes my research priorities. The country has 200 million people, fewer than 500 practicing psychiatrists, and no systematic environmental neurotoxin monitoring. My work addresses problems that matter here: addiction, environmental contamination, and the absence of computational tools for public health decision-making. The CCT model was motivated by the rising rate of substance use disorders among Nigerian youth. The environmental application was motivated by the Lagos Lagoon crisis, where industrial pollution has made fishing unsafe and drinking water unreliable.

I am applying to graduate programmes for October 2026 entry, with Medical University of Graz in Austria as my primary target. The EWRE grant would fund the computational development and field validation work between now and that enrollment. I am eligible for early-career, pre-PhD, and LMIC-track funding programmes. My long-term goal is to establish a computational environmental health laboratory in Lagos, training Nigerian scientists in AI-driven risk assessment and building tools that serve African communities.

CHECKLIST

- [ ] Identify a U.S. academic collaborator at University of Michigan or Georgia Tech to serve as PI and submitting organization
- [ ] Confirm collaborator's willingness to submit to NSF EWRE programme
- [ ] Verify collaborator's eligibility as NSF PI (U.S. institution, faculty or research scientist status)
- [ ] Draft collaborator letter of support describing their role and institutional commitment
- [ ] Prepare biosketch for Eniola Ayodele Olutogun (NSF format, 2-page limit)
- [ ] Prepare biosketch for collaborator
- [ ] Write project summary (1-page, NSF format)
- [ ] Write project description (15-page maximum, NSF format)
- [ ] Write budget justification with detailed cost breakdown
- [ ] Compile references cited (no page limit)
- [ ] Gather supporting documents: ORCID profile, GitHub repositories, preprint links, patent filing receipt
- [ ] Obtain letters of endorsement from Kent Berridge, Samuel Gershman, Nathaniel Daw, Marcelo Mattar
- [ ] Verify NSF EWRE deadline on programme website
- [ ] Confirm whether NSF requires U.S. citizenship or permanent residency for co-investigators
- [ ] Check if NSF allows foreign organizations as subawardees or if collaborator must manage all funds

EDITOR NOTES

- Eligibility risk: NSF EWRE typically requires the PI to be at a U.S. institution. Eniola cannot be PI. The collaborator must be confirmed before submission. If no collaborator is secured, this application cannot proceed.
- Fact to verify: Confirm that NSF allows foreign nationals as co-investigators on standard research grants. Some NSF programmes restrict co-PI eligibility to U.S. citizens or permanent residents. Check the NSF Proposal and Award Policies and Procedures Guide.
- Gap to fill: The profile does not specify which U.S. collaborator has been contacted or agreed to participate. This must be resolved before writing the full proposal. University of Michigan and Georgia Tech are suggested, but no confirmation exists.
- Budget detail needed: The 53,000 USD estimate is preliminary. NSF requires a detailed budget with specific cost categories, fringe benefits, indirect costs, and institutional facilities and administrative costs. The collaborator's institution will set the indirect cost rate.
- Missing personal detail: The profile does not include Eniola's date of birth, which may be required for NSF biographical sketches. Also missing: any prior grant funding, teaching experience, or awards that could strengthen the personal statement.

AI Draft — Energy, Water, and Resource Engineering