Draft — Women in Tech Accelerator

MOTIVATION LETTER

The eighth edition of the Women in Tech Accelerator targets early-stage, women-led technology ventures preparing to scale, with a five-month investment readiness program and a Demo Day competition for equity-free prizes. My venture, psyche-twin, fits that profile directly. It is a pre-seed AI company building a typed, multi-scale knowledge-graph architecture for self-modeling, and I am its founder and sole engineer. The program's focus on women entrepreneurs in technology, its emphasis on scalability and innovation, and its structured path to investor readiness match exactly where psyche-twin is now: a working prototype with a clear technical foundation and a defined path to product-market fit.

My background is pharmacist-turned-machine-learning-engineer. I hold a B.Pharm from the University of Ibadan (CGPA 5.1/7.0, German equivalent 1.9) and am enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute, University of Potsdam, starting Winter Semester 2026/27. I have built and validated four independent data pipelines using DuckDB for ingest-to-analyze corpus processing across life sciences, tech/AI, and social science domains. I self-host local LLM serving with llama.cpp and manage production Linux systems with CI/CD and automated backup. These are the operational skills required to run a lean AI startup.

psyche-twin's architecture is the differentiator. Multiple independent evidence streams, including LLM-derived, assessment-derived, behavioral, and document-derived data, fuse into one append-only event log. Disagreement between streams becomes an explicit graph edge rather than being averaged away. A first-person conversational interface generates responses grounded in the graph's current state. This design solves a real problem in AI self-modeling: most systems flatten conflicting evidence into a single score, losing the structure of how a person changes over time. psyche-twin preserves that structure.

The venture is pre-revenue and pre-product-market-fit. What exists is a typed knowledge-graph architecture, a working conversational interface, and a clear technical specification. The Women in Tech Accelerator's investment readiness training and Demo Day structure are the right next step to move from prototype to pilot users and initial funding. The program's $150,000 total prize pool at Demo Day is a concrete target, and the five-month timeline from June to October 2026 aligns with my current enrollment schedule.

I am a Nigerian woman building a technology venture with global application. The accelerator's support for women-led tech startups and its focus on economic diversification through innovation are directly relevant to my goals. I am prepared to commit to the full program duration and to the intensive work of preparing psyche-twin for investment.

RESEARCH STATEMENT

psyche-twin is a typed, multi-scale knowledge-graph architecture for self-modeling. The core technical problem is how to represent a human subject's psychological state over time when evidence streams disagree. Standard approaches average or discard conflicting signals. psyche-twin instead treats disagreement as first-class data: each evidence stream, whether LLM-derived, assessment-derived, behavioral, or document-derived, writes events to a single append-only log, and conflicts between streams become explicit graph edges with typed semantics. The system never erases a contradiction; it records it and reasons over it.

The architecture has three layers. The first is the event log, an append-only store that guarantees temporal integrity. The second is the typed knowledge graph, where nodes represent entities such as beliefs, behaviors, assessments, and external events, and edges carry type information including agreement, disagreement, causation, and temporal precedence. The third is the conversational interface, a first-person system that queries the graph's current state and generates responses grounded in that state, not in a generic language model prior.

I built the system myself. The backend uses Supabase/Postgres for the event log and graph storage. The frontend is JavaScript/Node.js. The LLM layer runs locally via llama.cpp with on-demand model swapping, which keeps data private and costs predictable. I have deployed the full stack on a Linux VPS with systemd, Caddy TLS, and automated backup and disaster recovery. The system passes its internal consistency tests: graph state is reproducible from the event log, and interface responses reflect the graph's current state rather than a static prompt.

The current validation status is honest and specific. psyche-twin has a working prototype and a complete technical specification. It does not yet have pilot users, revenue, or external validation of the self-modeling approach against standard psychological assessment instruments. The next phase is a pilot study with a small cohort, comparing graph-derived insights against established assessment tools to test whether the disagreement-preserving architecture produces measurably different, and clinically useful, outputs.

The Women in Tech Accelerator's investment readiness training is the correct vehicle for this stage. The program's emphasis on scalability applies directly: the architecture is domain-agnostic and can extend beyond self-modeling to any multi-stream evidence fusion problem. The Demo Day structure forces a clear investor narrative, which is precisely what psyche-twin needs to move from prototype to seed funding. I am applying with a technically sound prototype and a rigorous plan to validate it.

ESSAY: INNOVATION AND SCALABILITY

psyche-twin's innovation is architectural. Existing self-modeling and digital twin systems in health tech rely on a single data source or average multiple sources into a composite score. That design destroys information. My system preserves it by making disagreement an explicit, typed edge in a knowledge graph. This is a measurable technical difference, not a marketing claim.

The scalability argument is structural. The event log and graph schema are domain-agnostic. The same architecture that models a person's psychological state can model patient-reported outcomes in a clinical trial, user behavior in a digital health product, or any setting where multiple evidence streams must be reconciled without flattening. The append-only log design also creates a natural audit trail, which matters for regulated health applications.

The market context is concrete. Digital health and AI-driven assessment tools are growing sectors, and the specific problem of evidence fusion across heterogeneous data streams is unsolved in most commercial products. My background as a licensed pharmacist (PCN-licensed, B.Pharm from the University of Ibadan) and my current M.Sc. in Digital Health at the Hasso Plattner Institute give me both clinical credibility and technical depth. I have also built and validated four independent DuckDB-based data pipelines across life sciences, tech/AI, and social science domains, which means the data engineering foundation of psyche-twin is the same class of work I have already shipped.

The accelerator's focus on women-led tech startups with global impact potential fits my position. I am a Nigerian woman building a venture that addresses a universal problem in AI systems. The five-month program structure, from June to October 2026, is compatible with my M.Sc. enrollment, and the Demo Day competition provides a concrete milestone for investor readiness.

ESSAY: COMMITMENT AND PROGRAM FIT

The Women in Tech Accelerator's eighth edition runs a five-month program from June to October 2026, with investment readiness training and a Demo Day where participants compete for a total of $150,000 in equity-free cash prizes. I am prepared to commit to the full duration and to the program's requirements.

My current situation supports full participation. I am enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute, University of Potsdam, starting Winter Semester 2026/27, which begins after the accelerator's October 2026 end date. My current employment as National Product Manager at Synthcare, which began March 2026, is a role I can structure around the program's demands. The venture itself is at the pre-seed stage, which is the accelerator's stated target: early-stage startups preparing to scale.

The program's emphasis on economic diversification through innovation aligns with my goals. I am building a technology venture that can operate from Nigeria, from Germany, or from the UAE. The infrastructure is fully remote: the codebase lives on my own servers, the development workflow is asynchronous, and the product is a software service with no physical footprint. This means relocation or periodic presence in the UAE is technically feasible if the program requires it.

I am applying because the program's structure matches my venture's needs. psyche-twin has a working prototype but needs investor narrative development, pitch training, and exposure to a funding network. The accelerator provides all three. I am not seeking a biotech-specific program because psyche-twin is an AI infrastructure company with health applications, not a biotech company. The Women in Tech Accelerator's general technology focus is the correct fit.

CHECKLIST

- [ ] Complete online application form at the program URL (en.incarabia.com application page)
- [ ] Verify current application deadline on the program website
- [ ] Confirm UAE residency or visa status requirement; if required, assess relocation feasibility before submitting
- [ ] Prepare pitch deck (10-15 slides) covering psyche-twin architecture, validation status, and market opportunity
- [ ] Prepare one-page executive summary of psyche-twin
- [ ] Gather proof of enrollment for M.Sc. Digital Health at Hasso Plattner Institute (Winter Semester 2026/27)
- [ ] Update CV to include current employment (National Product Manager, Synthcare, March 2026-present)
- [ ] Prepare financial projection summary for pre-seed stage (burn rate, funding ask, use of funds)
- [ ] Identify and list two professional references (one technical, one clinical)
- [ ] Draft responses to any additional program-specific questions not covered in this document
- [ ] Submit application before deadline and save confirmation email

EDITOR NOTES

- Eligibility risk: The deep research notes indicate the program may require UAE-based founders. The applicant has no stated UAE ties. This is the single largest risk to the application. Confirm the residency requirement directly with the program before investing further time. If UAE presence is mandatory, assess whether the program allows remote participation or short-term relocation.
- Venture stage honesty: psyche-twin is pre-revenue, pre-product-market-fit, and has no pilot users. The application materials above state this explicitly. Do not let any reviewer or interviewer pressure the applicant into overstating validation status. The technical prototype is real and tested; the commercial validation is not.
- Verification needed: Confirm the program's exact start date (stated as June 2026 in the profile) and end date (October 2026) against the current program website. Confirm the $150,000 prize pool figure and whether it is cash, equity-free, or a combination. Confirm whether the program accepts applicants who are enrolled in a degree program concurrently.
- Personal detail gap: The applicant's location as of the application date is not stated in the profile. If the applicant is in Nigeria, Germany, or elsewhere, this affects the UAE residency question and should be addressed in the application form directly.
- Selection of psyche-twin over other research lines: The CCT model, TOPOLOGIX, and neurocascade are research projects with journal submissions, not ventures. The Women in Tech Accelerator is a startup accelerator, not a research fellowship. psyche-twin is the only line in the profile that is a venture with a product architecture and a path to scale. The other lines are cited for credibility in the applicant's background, not presented as the venture itself.

MOTIVATION LETTER

The eighth edition of the Women in Tech Accelerator targets early-stage, women-led technology ventures preparing to scale, with a five-month investment readiness program and a Demo Day competition for equity-free prizes. My venture, psyche-twin, fits that profile directly. It is a pre-seed AI company building a typed, multi-scale knowledge-graph architecture for self-modeling, and I am its founder and sole engineer. The program's focus on women entrepreneurs in technology, its emphasis on scalability and innovation, and its structured path to investor readiness match exactly where psyche-twin is now: a working prototype with a clear technical foundation and a defined path to product-market fit.

My background is pharmacist-turned-machine-learning-engineer. I hold a B.Pharm from the University of Ibadan (CGPA 5.1/7.0, German equivalent 1.9) and am enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute, University of Potsdam, starting Winter Semester 2026/27. I have built and validated four independent data pipelines using DuckDB for ingest-to-analyze corpus processing across life sciences, tech/AI, and social science domains. I self-host local LLM serving with llama.cpp and manage production Linux systems with CI/CD and automated backup. These are the operational skills required to run a lean AI startup.

psyche-twin's architecture is the differentiator. Multiple independent evidence streams, including LLM-derived, assessment-derived, behavioral, and document-derived data, fuse into one append-only event log. Disagreement between streams becomes an explicit graph edge rather than being averaged away. A first-person conversational interface generates responses grounded in the graph's current state. This design solves a real problem in AI self-modeling: most systems flatten conflicting evidence into a single score, losing the structure of how a person changes over time. psyche-twin preserves that structure.

The venture is pre-revenue and pre-product-market-fit. What exists is a typed knowledge-graph architecture, a working conversational interface, and a clear technical specification. The Women in Tech Accelerator's investment readiness training and Demo Day structure are the right next step to move from prototype to pilot users and initial funding. The program's $150,000 total prize pool at Demo Day is a concrete target, and the five-month timeline from June to October 2026 aligns with my current enrollment schedule.

I am a Nigerian woman building a technology venture with global application. The accelerator's support for women-led tech startups and its focus on economic diversification through innovation are directly relevant to my goals. I am prepared to commit to the full program duration and to the intensive work of preparing psyche-twin for investment.

RESEARCH STATEMENT

psyche-twin is a typed, multi-scale knowledge-graph architecture for self-modeling. The core technical problem is how to represent a human subject's psychological state over time when evidence streams disagree. Standard approaches average or discard conflicting signals. psyche-twin instead treats disagreement as first-class data: each evidence stream, whether LLM-derived, assessment-derived, behavioral, or document-derived, writes events to a single append-only log, and conflicts between streams become explicit graph edges with typed semantics. The system never erases a contradiction; it records it and reasons over it.

The architecture has three layers. The first is the event log, an append-only store that guarantees temporal integrity. The second is the typed knowledge graph, where nodes represent entities such as beliefs, behaviors, assessments, and external events, and edges carry type information including agreement, disagreement, causation, and temporal precedence. The third is the conversational interface, a first-person system that queries the graph's current state and generates responses grounded in that state, not in a generic language model prior.

I built the system myself. The backend uses Supabase/Postgres for the event log and graph storage. The frontend is JavaScript/Node.js. The LLM layer runs locally via llama.cpp with on-demand model swapping, which keeps data private and costs predictable. I have deployed the full stack on a Linux VPS with systemd, Caddy TLS, and automated backup and disaster recovery. The system passes its internal consistency tests: graph state is reproducible from the event log, and interface responses reflect the graph's current state rather than a static prompt.

The current validation status is honest and specific. psyche-twin has a working prototype and a complete technical specification. It does not yet have pilot users, revenue, or external validation of the self-modeling approach against standard psychological assessment instruments. The next phase is a pilot study with a small cohort, comparing graph-derived insights against established assessment tools to test whether the disagreement-preserving architecture produces measurably different, and clinically useful, outputs.

The Women in Tech Accelerator's investment readiness training is the correct vehicle for this stage. The program's emphasis on scalability applies directly: the architecture is domain-agnostic and can extend beyond self-modeling to any multi-stream evidence fusion problem. The Demo Day structure forces a clear investor narrative, which is precisely what psyche-twin needs to move from prototype to seed funding. I am applying with a technically sound prototype and a rigorous plan to validate it.

ESSAY: INNOVATION AND SCALABILITY

psyche-twin's innovation is architectural. Existing self-modeling and digital twin systems in health tech rely on a single data source or average multiple sources into a composite score. That design destroys information. My system preserves it by making disagreement an explicit, typed edge in a knowledge graph. This is a measurable technical difference, not a marketing claim.

The scalability argument is structural. The event log and graph schema are domain-agnostic. The same architecture that models a person's psychological state can model patient-reported outcomes in a clinical trial, user behavior in a digital health product, or any setting where multiple evidence streams must be reconciled without flattening. The append-only log design also creates a natural audit trail, which matters for regulated health applications.

The market context is concrete. Digital health and AI-driven assessment tools are growing sectors, and the specific problem of evidence fusion across heterogeneous data streams is unsolved in most commercial products. My background as a licensed pharmacist (PCN-licensed, B.Pharm from the University of Ibadan) and my current M.Sc. in Digital Health at the Hasso Plattner Institute give me both clinical credibility and technical depth. I have also built and validated four independent DuckDB-based data pipelines across life sciences, tech/AI, and social science domains, which means the data engineering foundation of psyche-twin is the same class of work I have already shipped.

The accelerator's focus on women-led tech startups with global impact potential fits my position. I am a Nigerian woman building a venture that addresses a universal problem in AI systems. The five-month program structure, from June to October 2026, is compatible with my M.Sc. enrollment, and the Demo Day competition provides a concrete milestone for investor readiness.

ESSAY: COMMITMENT AND PROGRAM FIT

The Women in Tech Accelerator's eighth edition runs a five-month program from June to October 2026, with investment readiness training and a Demo Day where participants compete for a total of $150,000 in equity-free cash prizes. I am prepared to commit to the full duration and to the program's requirements.

My current situation supports full participation. I am enrolled in the M.Sc. Digital Health program at the Hasso Plattner Institute, University of Potsdam, starting Winter Semester 2026/27, which begins after the accelerator's October 2026 end date. My current employment as National Product Manager at Synthcare, which began March 2026, is a role I can structure around the program's demands. The venture itself is at the pre-seed stage, which is the accelerator's stated target: early-stage startups preparing to scale.

The program's emphasis on economic diversification through innovation aligns with my goals. I am building a technology venture that can operate from Nigeria, from Germany, or from the UAE. The infrastructure is fully remote: the codebase lives on my own servers, the development workflow is asynchronous, and the product is a software service with no physical footprint. This means relocation or periodic presence in the UAE is technically feasible if the program requires it.

I am applying because the program's structure matches my venture's needs. psyche-twin has a working prototype but needs investor narrative development, pitch training, and exposure to a funding network. The accelerator provides all three. I am not seeking a biotech-specific program because psyche-twin is an AI infrastructure company with health applications, not a biotech company. The Women in Tech Accelerator's general technology focus is the correct fit.

CHECKLIST

- [ ] Complete online application form at the program URL (en.incarabia.com application page)
- [ ] Verify current application deadline on the program website
- [ ] Confirm UAE residency or visa status requirement; if required, assess relocation feasibility before submitting
- [ ] Prepare pitch deck (10-15 slides) covering psyche-twin architecture, validation status, and market opportunity
- [ ] Prepare one-page executive summary of psyche-twin
- [ ] Gather proof of enrollment for M.Sc. Digital Health at Hasso Plattner Institute (Winter Semester 2026/27)
- [ ] Update CV to include current employment (National Product Manager, Synthcare, March 2026-present)
- [ ] Prepare financial projection summary for pre-seed stage (burn rate, funding ask, use of funds)
- [ ] Identify and list two professional references (one technical, one clinical)
- [ ] Draft responses to any additional program-specific questions not covered in this document
- [ ] Submit application before deadline and save confirmation email

EDITOR NOTES

- Eligibility risk: The deep research notes indicate the program may require UAE-based founders. The applicant has no stated UAE ties. This is the single largest risk to the application. Confirm the residency requirement directly with the program before investing further time. If UAE presence is mandatory, assess whether the program allows remote participation or short-term relocation.
- Venture stage honesty: psyche-twin is pre-revenue, pre-product-market-fit, and has no pilot users. The application materials above state this explicitly. Do not let any reviewer or interviewer pressure the applicant into overstating validation status. The technical prototype is real and tested; the commercial validation is not.
- Verification needed: Confirm the program's exact start date (stated as June 2026 in the profile) and end date (October 2026) against the current program website. Confirm the $150,000 prize pool figure and whether it is cash, equity-free, or a combination. Confirm whether the program accepts applicants who are enrolled in a degree program concurrently.
- Personal detail gap: The applicant's location as of the application date is not stated in the profile. If the applicant is in Nigeria, Germany, or elsewhere, this affects the UAE residency question and should be addressed in the application form directly.
- Selection of psyche-twin over other research lines: The CCT model, TOPOLOGIX, and neurocascade are research projects with journal submissions, not ventures. The Women in Tech Accelerator is a startup accelerator, not a research fellowship. psyche-twin is the only line in the profile that is a venture with a product architecture and a path to scale. The other lines are cited for credibility in the applicant's background, not presented as the venture itself.

AI Draft — Women in Tech Accelerator