Grant and cloud partner brief

12-month runway to prove an end-to-end prevention stack.

Live today: Teenology PWA and a working specialist web prototype. Grant cloud credits would fund engineering completion, production inference, and a bounded pilot—not a claim of national deployment.

Try fund demo Specialist prototype

How we would use partner credits

Two clouds, two jobs

Credits are not interchangeable marketing labels—they map to concrete cost lines in a solo-founder build.

Microsoft for Startups · Azure

Build, host, and harden

Production infrastructure: APIs, staging, PostgreSQL direction, safety-test environments
Azure OpenAI: specialist-desk mediation tests, multi-step document flows, automated safety regression runs
Engineering velocity: cloud budget for AI-assisted development in the IDE (e.g. Cursor) so one founder can ship federation, terminals, and compliance hooks faster—this is a development line item, not family-facing chat tokens

Google for Startups · Cloud

Run Teenology inference at scale

Gemini / Vertex: long-context Family Bridge, diary continuity, and consumer companion sessions—the ongoing token cost of live family use
Optional Ads credits: small, measured pilot to learn if international positioning resonates (no CAC targets published until a baseline exists)
Analytics contour: BigQuery research pipeline for anonymized pilot metrics (design described on Science)

12-month grant plan

Timeline, KPIs, and credit use

Figures below are planning targets for fund reviewers; final numbers follow the application form.

Phase map (months 0–12)

Phase	Focus	Exit criteria
0–3	L1↔L2 consent bridge, safety audit, specialist web polish	Documented crisis-routing tests; demo path for reviewers
4–6	Bounded family pilot (fund-demo cohort)	≥50 active families with check-up + 2+ sessions; Pulse/Battery telemetry
7–9	Specialist workspace beta (school psychologists)	≥15 specialists generating .docx drafts with PIE tags
10–12	Aggregate dashboard prototype (synthetic + pilot aggregates)	One regional-style sandbox dashboard; privacy review memo

KPIs reviewers can verify

Live URLs: teenology.care, web.prevention.school, prevention.school
Founder: ORCID 0000-0001-9876-9798; contracting entity — IE Roman Dubrovskii, Tbilisi, ID 302269747 (registry extract supplied with application)
Safety: documented Safety Loop triggers + human hotline handoff (see Teenology ethics)
Pilot engagement: weekly active families, check-up completion rate, crisis-route activations (counts only)
Specialist value: time-to-first usable .docx draft; taxonomy tag coverage
No claim of clinical efficacy until formal study design—process and architecture metrics first

Where credits go (budget buckets)

~40% Google — consumer inference tokens (Gemini) during pilot
~35% Azure — hosting, databases, Azure OpenAI test workloads
~15% Azure — engineering tooling & CI safety automation
~10% compliance copy, privacy review, pilot instrumentation

Percentages are illustrative for credit allocation, not a cash budget. Cash cofunding, if any, is listed in the formal application.

Privacy & compliance

Privacy-by-design (not a compliance guarantee)

The architecture separates personal identifiable information (PII) from the LLM processing layer. Consumer diaries can stay local-first; the design targets GDPR- and COPPA-aligned deployment—final compliance depends on jurisdiction, DPIAs, and institutional contracts.

What funding unlocks

Deliverables at month 12

Connected L1–L2 demo

Teenology prevention signals flow to specialist workspace only with explicit consent—showable to ministries and schools.

Specialist-ready workspace

Secure web terminal with Consultant / Expert / Architect modes and exportable prevention documents.

Dashboard blueprint + pilot aggregates

Regional-style analytics UI fed by anonymized pilot data—not a live national registry.