Skip to content

Industries · Healthcare

Healthcare

Silex designs, supplies, and operates the infrastructure, security, recovery, and AI governance that health systems and research institutions depend on.

Where we focus

  • Clinical system recovery
  • PHI protection at the AI gateway
  • HIPAA-aligned operations
  • Research institution automation

How we fit

Where Silex works in a health system.

A health system runs clinical, research, and business systems on one set of infrastructure. Silex works at four points in that stack, from the AI gateway down to the recovery environment.

HIPAA-aligned operations across every layer
HIPAA-aligned operationsSecurity controls from Zero Trust access to security analytics, and Forward Deployed Engineers who operate the platforms with your team.Cybersecurity →Managed services →
Layer 01PeopleClinicians, researchers, and administrative staff.
Layer 02ApplicationsClinical, research, and business applications, and new AI assistants.
Layer 03AI and integrationModel calls, agent tool calls, and the APIs that connect systems.
PHI protection at the AI gatewayPrompts and responses are redacted and audited before they reach any model.AI and MCP Gateway →
Layer 04PlatformsData centers, storage, cloud, and the automation that runs them.
Research institution automationAutomation programs on Ansible Automation Platform and OpenShift for research environments.Automation →
Layer 05RecoveryImmutable copies and an isolated environment to recover into.
Clinical system recoveryIsolated recovery environments that bring clinical systems back after an attack.Cyber resilience →

Healthcare work

Recovery and AI governance for health systems.

Silex designs and operates isolated recovery environments for healthcare, tested through structured recovery exercises.

Recovered systems pass through four isolation zones and five recovery phases before they reconnect, and every test ends with an after-action review.

Silex implements Kong AI Gateway and MCP governance for multiple customers and runs it as a managed service.

One governed endpoint carries model calls and agent tool calls, with redaction, token budgets, and an audit trail for every request.

Clinical documentation AI

From our own model development.

Silex engineers built and evaluated models that classify evaluation and management levels and flag documentation gaps in clinical notes. Real patient data could not be used for training, so the models were trained on synthetic data from a multi-stage generation pipeline.

This was Silex research and development work. It is the experience we bring to clinical AI use cases, not a product we sell.

Applied AI Engineering
01 · DataSynthetic clinical notesPlanned, generated, re-scored by a second model, and labeled against a closed vocabulary.
02 · ModelClassificationFine-tuned open-weight models, with coding rules kept in code alongside the model.
03 · OutputLevel and gapsA suggested evaluation and management level, with documentation gaps flagged for a person to review.

Frameworks

Mapped to the frameworks your auditors use.

  • HIPAA Security Rule safeguards, including PHI in AI prompts and responses
  • HIPAA contingency planning, including testing and revision of the plan
  • NIST Cybersecurity Framework 2.0 Recover function
  • NIST AI Risk Management Framework for AI governance

Platforms

What we build with in healthcare.

Platinum Partner

Certified Delivery Partner

Premier Partner

Premier Partner

Cohesity · Commvault · Zerto · VM2020 · Zscaler · Palo Alto Networks · CrowdStrike

Questions buyers ask

Questions

How do you keep PHI out of AI prompts?

The AI gateway sits between your applications and every model. It redacts PHI from prompts before they leave your environment, applies guardrails to responses, and logs each request for audit. The gateway data plane runs in your infrastructure.

How does a hospital recover after ransomware?

An isolated recovery environment keeps immutable copies of clinical and business systems out of production's reach. Systems are restored into an air-gapped zone, cleaned, validated, and reconnected in a set order, and staff use recovered applications through virtual desktops during the incident.

How do we test our HIPAA contingency plan?

Silex runs recovery tests in tiers, from quick smoke tests to full incident simulations, and holds an after-action review after each one. The test records and reviews show auditors that the plan is tested and revised.

Do you work with research institutions?

Yes. Silex builds and operates automation programs on Ansible Automation Platform and OpenShift for research environments, and designs GPU environments for research and AI workloads.

Where should we start?

Most healthcare engagements start with a Cyber Recovery Maturity Assessment, an AI Governance Readiness Review, or a ransomware recovery workshop. Each one ends with a roadmap your team can act on.

Next step

Talk to a healthcare engineer.

Tell us where recovery, security, or AI governance concerns you most, and we will recommend where to start.