Frequently Asked Questions
Answers to common questions about SynthVault, our vault platform, and how we help organisations govern AI model assets.
SynthVault is a secure AI model vault platform — not a marketing agency, advertising service, or creative consultancy. The questions below address what we do, how our vault works, and what you can expect when working with our team in Vancouver.
Is SynthVault a marketing agency?
No. SynthVault is exclusively a secure AI model vault platform. We provide technology infrastructure for storing, governing, versioning, and deploying machine learning assets. We do not offer advertising services, brand strategy, content marketing, social media management, search engine marketing, lead generation campaigns, or any form of promotional creative work. If you are looking for marketing services, SynthVault is not the right partner. If you need to secure and govern your AI models with enterprise-grade controls, we are built for exactly that purpose.
What types of AI assets can be stored in the vault?
SynthVault supports model weights (PyTorch, TensorFlow, ONNX, Hugging Face formats), training and evaluation datasets, prompt libraries, evaluation suites, deployment manifests, container images, and configuration files. Each asset type has dedicated metadata schemas, version tracking, and access control policies.
How does SynthVault handle Canadian privacy law?
Our platform is designed with PIPEDA compliance as a core requirement. We provide consent tracking for datasets containing personal information, data residency configuration for Canadian storage, data subject access request workflows, and audit logs that support privacy impact assessments.
Can SynthVault integrate with our existing ML toolchain?
Yes. We support integrations with MLflow, Kubeflow, SageMaker, Azure ML, Vertex AI, GitHub Actions, Docker, Kubernetes, and custom CI/CD pipelines. Our integration engineering team can build connectors for proprietary systems as part of a professional services engagement.
What deployment options are available?
SynthVault can be deployed as a managed cloud service, in your private cloud environment, or fully on-premises. Hybrid configurations are also supported for organisations that need certain model assets to remain on-site while using cloud resources for others.
Do you guarantee zero security breaches?
No. No technology platform can honestly guarantee zero security breaches. SynthVault implements defence-in-depth security including AES-256 encryption at rest, TLS 1.3 in transit, role-based access controls, multi-factor authentication, anomaly detection, and regular third-party penetration testing. However, we acknowledge that determined adversaries, zero-day vulnerabilities, and human error can create security incidents. Our commitment is to minimise risk through robust controls, maintain transparent incident response procedures, and notify affected clients promptly if a breach occurs. We believe honesty about security limitations builds more trust than false guarantees.
Do you guarantee AI model accuracy or business outcomes?
No. SynthVault secures and governs your AI model assets — we do not develop, train, or tune models on your behalf. The accuracy, fairness, reliability, and business impact of AI systems built on vault-stored assets remain entirely the responsibility of your data science and engineering teams. Our platform provides the infrastructure for version control, access management, and deployment governance, not AI performance optimisation.
How long does vault onboarding typically take?
Our Vault Foundations program completes initial onboarding in six weeks, including migration of your first ten models. Enterprise deployments with complex integration requirements may take twelve to sixteen weeks. Timelines depend on the size of your existing model portfolio and the number of systems requiring integration.
What support is included?
All clients receive email support during business hours (Pacific Time). Enterprise clients receive priority support with defined response time SLAs, a dedicated vault architect, and access to our 24/7 managed operations service for production vault instances.
How do I request a demo?
Visit our contact page and submit the enquiry form. Our team will schedule a personalised vault demonstration, typically within three business days. Demos are conducted remotely or on-site at our Vancouver office.