The Cytognosis infrastructure follows extreme simplicity, serverless-first, and HIPAA-ready architectural choices. This repository contains the source of truth for our deployment strategies, domain mappings, identity federation, data analytics ecosystems, and the data strategy and scholarly knowledge graph that everything serves.
Topology for the cytognosis- GCP projects, the historical context of the cg- and *-zone dual Cloud DNS zones, .org / .com / .ai management, third-party domain validations (Anthropic), and the Serverless IP scheme.
Migration off legacy GCS static buckets into fully containerized Google Cloud Run deployments. GitHub Actions OIDC federation (website-deployer), Uvicorn debugging, and the artifact-registry layout.
Nonprofit Google Workspace environment, automated deprecation of Google Apps Scripts, retirement of the noreply@cytognosis.org paid seat, and the inbound alias filtering network (careers@, partnerships@, science@, info@, stories@, gcp-data-scientists@, gcp-infra-admins@).
Unified data hub. Master strategy, public dataset strategy, dataset catalog, FAIRification mandate, HIPAA program, controlled-access SOPs (NIH NDA + academic partnerships), multimodal schemas, DUA template — plus the scholarly knowledge graph stack:
Deep dive into the GCS bucket taxonomy, VPC Service Controls, Cloud Healthcare API, Confidential Compute, and IAM controls for sensitive genomics and neuroscience datasets.
Single source of truth for self-hosted operations. Architecture, evaluation reports, and deployment walkthroughs for Caddy, Cal.com, Excalidraw, Mermaid Live, Logseq, MLflow, Neo4j, and Jupyter — all powered by the internal Container Framework.
Comprehensive 911-tool technology landscape across the Cytognosis 16-layer infrastructure stack (L1-storage through L16-cloud). Three product surfaces (Cytonome, Cytoverse, Spine), five adoption statuses, seven cross-cutting invariants (open-by-default, schema-as-hub, local-first, MCP interop, operational FAIR, dual-tier S2S, reproducibility). Includes the recommended stack at a glance, voice architecture options, storage hierarchy by data type, neuroimaging/connectomics stack, and full rejection/deferral rationale.
Active scientific programs with dedicated data environments, regulatory requirements, and external collaborators.
Cytognosis's first major research program: a multimodal, transdiagnostic, individual-level
foundation model for brain and psyche. Collaborators: McLean Hospital / MGB (Brad Ruzicka)
and Purdue University (Ananth Grama).
Six-principle reproducibility framework applying to all Cytognosis computational outputs
(cytoskeleton, cytos, Yar, neuro-*). Goal: any published result re-runnable with a singlecytoskeleton reproduce by a third party.
[!NOTE]
HIPAA compliance dashboard: data-strategy/compliance/HIPAA-STATUS.md —
single pane of glass for all 45 CFR controls with current status, evidence links, and owners.>
NIH GDS 2025 requirements: data-strategy/compliance/nih-gds-requirements.md —
8 operational requirements, NIST SP 800-171 mapping, generative AI restriction, DUC checklist.