MedivistyΛAINNA NeuralOps
Explore partnership ↗
STRATEGIC TECHNICAL PARTNERSHIP · 01 / 12

Healthcare innovation,
engineered with AI.

Medivisty × AINNA NeuralOps. An AI-native execution platform for faster development, secure healthcare infrastructure, real-time modifications and sustainable operating costs.

Explore the proposal →View commercial model ↗Download PowerPoint ↓
↗
Build fasterRequirements to live systems
◎
Lower overheadSmarter AI and infrastructure spend
♢
Secure by designWAF, audit, access and recovery
⟳
Change continuouslyControlled updates and optimization
02 / Platform overview

One execution layer.
The entire lifecycle.

NeuralOps is positioned as a horizontal AI orchestration platform across software development, infrastructure, security and operations — with Medivisty’s healthcare experts retaining domain ownership.

02 / 12
Operating principle: Medivisty defines clinical, educational and business requirements. NeuralOps coordinates technical execution with testing, approvals and auditability.
View slide reference · Introduction to NeuralOpsNeuralOps introduction slide
03 / Problem statement

Where complexity
becomes expensive.

These are potential challenges to validate during discovery, not findings about Medivisty’s current systems.

03 / 12
04 / Reference architecture

Modular. Secure.
Built to evolve.

A proposed technical foundation linking healthcare applications, centralized data services, secure hosting and an AI-managed operations layer.

04 / 12

Users & Channels

Clinicians
Medical students
Operations teams
Patients / Public
Web · Mobile · PWA
AINNA NeuralOps · AI Agent Execution LayerPlan → Build → Test → Deploy → Operate → Optimize

Application modules

EHR / EMRMedical LMSAppointmentClinical workflowsLab resultsREST APIs

Application services

Laravel / PHP · Workers · API services · AI connector

Data services

PostgreSQLRedisObject StorageSAN / Backups

Infrastructure & Security

BunkerWeb WAFReverse proxy · ModSecurity · OWASP CRS
Docker / Compose
Linux · Firewall · SSH
RBAC · Secrets · Logs
Metrics · Backup · DR
CI/CD · Approval Gates
Reference architecture only. Final design, sizing and healthcare data controls depend on Medivisty’s actual workloads and regulatory requirements.
View detailed architecture diagramReference architecture overview
05 / Server engineering

From an empty server
to controlled production.

Standardize provisioning, OS hardening, deployment and observability instead of repeatedly configuring infrastructure by hand.

05 / 12
01 · ProvisionCloud, VPS, on-premise
→
02 · Install OSUbuntu LTS / Debian
→
03 · ContainerizeDocker / Compose
→
04 · HardenSSH · UFW · Fail2ban
→
05 · OperatePatches · Metrics · Backup

OS & Security Baseline

  • Least-privilege access and non-root operation
  • SSH key authentication and network segmentation
  • UFW firewall and Fail2ban / CrowdSec where appropriate
  • Security updates, vulnerability assessment and host logs

Operations & Reliability

  • Environment isolation: dev, staging, production
  • CPU, RAM, disk, network and service monitoring
  • Configuration management and rollback
  • Encrypted backups and tested restoration
Server sizing is determined by measured load, concurrency, storage needs, resilience objectives and expected growth — not fixed hardware assumptions.
06 / Web hosting & system management

A resilient application stack,
including SAN storage.

Containerized hosting and centralized systems management, with configurable SAN storage for workloads that justify shared high-availability storage.

06 / 12
BunkerWebWAF · Reverse proxy
→
Laravel / PHPWeb · APIs · Workers
↔
PostgreSQLRelational data
↔
RedisCache · Queue
↔
TrueNAS / OpenZFSiSCSI · SMB/NFS · Snapshots
▣

System Management

RBAC, asset inventory, system configuration, service health, patch management and operational audit logs.

▤
SAN storage

TrueNAS CORE / SCALE · OpenZFS

Built on OpenZFS, TrueNAS can provide block-level storage through iSCSI targets alongside traditional file sharing such as SMB and NFS. Snapshot and replication features support recovery and data movement.

TrueNASOpenZFSiSCSISMB / NFSSnapshotsReplication
⟳

Backup & Disaster Recovery

Encrypted, versioned backups with offsite copies; scheduled restoration tests and documented RTO/RPO targets.

TrueNAS CORE / SCALE and OpenZFS are a reference option, not a fixed design. Final storage choice depends on workload, network throughput, redundancy, capacity and support requirements. Snapshots and replication complement—not replace—independent backups and tested restores.
View detailed hosting stack referenceWeb hosting architecture and system management slide
07 / Defense-in-depth

Security starts
before the application.

BunkerWeb WAF provides an application-edge protection layer alongside host hardening, identity controls, monitoring and recovery processes.

07 / 12
⬡

BunkerWeb WAF

NGINX-based reverse proxy, ModSecurity and OWASP Core Rule Set for filtering common web attacks.

⊘

Traffic Protection

Rate limiting, bot mitigation, IP policies and request filtering, tuned to minimize false positives.

⚿

Identity & Secrets

RBAC, least privilege, MFA where appropriate, secret rotation and controlled access to sensitive data.

▦

Host & Container

UFW, intrusion prevention, patched images, vulnerability scans and network isolation.

☷

Audit & Visibility

Centralized logs, request tracing, security events, access review and incident-response workflows.

↻

Healthcare Governance

Data minimization, encryption, backup and policy-aligned handling of sensitive healthcare information.

A WAF does not by itself provide regulatory compliance or make an application immune to security vulnerabilities; additional validation and governance are required.
08 / AI efficiency

The right model.
For the right task.

Smart routing avoids running every request through an expensive frontier model. Route by task complexity, quality requirements, latency, privacy and cost.

08 / 12
Tier 1 · LightweightRoutine workflows

Classifications, simple transformations, standard responses, low-risk extraction and triage.

Tier 2 · BalancedDevelopment & analysis

Code changes, debugging, documentation, database transformations and controlled tool use.

Tier 3 · AdvancedComplex reasoning

Architecture, multi-system diagnosis, difficult planning and high-complexity tasks requiring stronger capabilities.

Task + ContextRequest · Risk · Data class
→
Policy GatewayPrivacy · Budget · Limits
→
Smart RouterQuality · Cost · Latency
→
AI AgentExecute · Test · Report
→
Approval GateFor sensitive changes
Safeguards: token budgets, caching, provider fallback, evaluation metrics, audit trails, escalation rules and human approval for production or clinical-risk changes.
09 / Managed operations

Measure everything.
Change with confidence.

Unified observability and disciplined release practices reduce the time between detecting an issue and applying a validated fix.

09 / 12
⌁

Observability

Prometheus metrics, Grafana dashboards and Loki or equivalent centralized log aggregation.

⚑

Incident Response

Health checks, severity levels, alert routing, runbooks, escalation and issue history.

↻

Safe Releases

Git-controlled changes, build pipelines, automated tests, approvals and rollback strategy.

▣

Data Recovery

Database and file backups, retention policies, restore tests and recovery objectives.

◎

Cost Visibility

Per-service resources, hosting spend, model token usage and allocation by project.

◈

Optimization

Rightsize infrastructure and identify expensive endpoints, queries, queues or model tasks.

10 / Healthcare domain enablement

Medical expertise leads.
AI accelerates execution.

Medivisty’s clinicians, educators and operational experts shape the products. NeuralOps supports implementation without taking ownership of clinical judgment or patient safety decisions.

10 / 12
Medivisty owns

Domain authority

  • Clinical requirements and validation
  • Medical education workflows and curriculum
  • Product vision, roadmap and acceptance criteria
  • Healthcare governance and patient safety decisions
NeuralOps enables

Technical execution

  • Solution design and AI-assisted software development
  • Infrastructure configuration and deployment
  • Automated testing, monitoring and maintenance
  • Documentation, change control and optimization
For clinically consequential functions, require clinician review, validated algorithms, appropriate privacy controls and explicit sign-off before use in patient-facing production environments.
11 / Commercial proposal

Predictable baseline.
Scales with project value.

A simple annual commercial model covering the agreed development, server-management and AI-agent scope, with a minimum annual fee.

11 / 12

Annual NeuralOps Service Fee

RM 20,000Minimum annual fee
OR
10%Of total project cost

Whichever is higher. Formula: Annual Fee = MAX(RM20,000, 10% × Project Cost).

Proposed service coverage

  • AI development and implementation support
  • Server provisioning, configuration and operations
  • AI agent orchestration and smart routing setup
  • Deployment, security, monitoring and maintenance
  • Continuous optimization and changes within agreed scope

Interactive cost estimator

RM
Estimated annual NeuralOps fee
RM 20,000
Minimum annual fee applies
Project valueAnnual fee
RM 100,000RM 20,000
RM 200,000RM 20,000
RM 300,000RM 30,000
RM 500,000RM 50,000
Commercial clarification: The fee is a proposal, not an executed contract. Define project valuation, included development hours/changes, API/model usage, cloud infrastructure bills, SAN hardware, applicable taxes, support SLA and third-party subscriptions in the final agreement.
12 / Partnership roadmap

Start with evidence.
Scale with confidence.

A controlled proof of concept on one selected Medivisty application validates the operational, technical and financial case before expansion.

12 / 12
01

Select a pilot

Choose an existing, bounded system or feature and define stakeholders, security scope and baseline metrics.

02

Assess the baseline

Review code, infrastructure, dependencies, current change process, hosting bills and AI usage.

03

Implement & verify

Apply NeuralOps-assisted development, hardening, observability and deployment under approval gates.

04

Evaluate & decide

Compare speed, reliability, security, costs and ongoing operational complexity with agreed KPIs.

Suggested KPIs: deployment lead time, change turnaround, escaped defects, uptime, incident resolution time, server utilization, infrastructure spend and AI cost per successful task.
13 / The experience behind the operating model

Built from hands-on work.
Extended with AI agents.

NeuralOps formalizes practical, end-to-end IT delivery: from infrastructure and applications to operations, with domain experts leading requirements and human approval retained for critical decisions.

SUPPLEMENTARY
AUCMS · First-person context

From building systems from the ground up

At AUCMS, I built and managed much of the IT environment myself: staff networking, pfSense, Snort IDS, Wi-Fi, routers, firewall and IDS/IPS, the AUCMS portal, and a Document Management System for A&R—including application development, server configuration and deployment.

I also developed and managed a Library Management System and LMS covering thousands of e-books. The library system ran for almost three years before UNITAR introduced its own system. Most of this work was handled solo, with IT colleagues assisting on PCs and selected components; I also managed company social media and photography.

Before AI agents, I sometimes had to remote in to troubleshoot server and network issues while away in Terengganu or Genting Highlands. That experience made one risk clear: reliable systems can still depend too heavily on a single person.

How NeuralOps applies it

Augment the team. Keep people in control.

  • Domain experts lead: Medivisty’s healthcare and medical-education specialists define workflows, requirements and acceptance.
  • Agents execute: help translate requirements into specifications, code, database structures, tests, integrations, deployment and maintenance.
  • Detached System: AI can help create scripts, parsers and automations; once tested and approved, repeatable work can run deterministically without an AI call each time.
  • Smart Model Routing: route simpler tasks to economical models and use higher-capability models only when justified.
  • Human gates remain: clinical validation, sensitive-data protection and production approval stay with accountable people.
Experience & technology ecosystem
Protocols & systemsiSCSISMBNFSIDS / IPSLibrary systemLMSDMSAUCMS experience

Official AINNA and Medivisty assets, TrueNAS and BunkerWeb brand assets, and product icons from Simple Icons (CC0). All marks remain the property of their respective owners.

Partnership principle: Medivisty provides healthcare domain expertise. AINNA NeuralOps supports technical execution, infrastructure and continuous improvement—without replacing developers, IT staff or clinical decision-makers.
PARTNERSHIP DISCUSSION

Healthcare expertise.
AI-native execution.

Medivisty leads the domain. AINNA NeuralOps supports secure, cost-aware delivery and infrastructure operations. Begin with one measurable pilot.

Review commercial model ↑Revisit architecture ↑
AINNA
CLICK ME

Site Sections

No section data available yet.

Sites with documented sections will appear here.