Scaleout Systems AB

Scaleout Edge Platform

Scaleout Edge is a platform for operational AI in demanding, distributed environments. It supports AI teams in collaboratively building, deploying, and governing machine learning models across edge, cloud, and secure domains using federated learning.

Features

• Federated Learning: Train models at the edge
• Telemetry & Monitoring: Know exactly how models perform
• Secure Model Propagation: Update models over encrypted channels
• Orchestration: Control distributed AI from a single pane
• Model Development: Turn any ML code into edge-deployable packages
• ModelOps
• Provenance Tracking: Explainable evidence of everything that happened
• Vision Module: Solution enabler for computer vision
• UAV kit: Solution enabler for drones
• Model robustness and resilience testing toolkit

Benefits

• Enable AI on distributed, isolated datasets
• Enable AI without moving raw data from original location
• Enable AI inference and training in low-connectivity or disconnected environments
• Allow for continuously updated AI on edge
• Privacy Enhancing Technology

Service documents

• Pricing document

  PDF

• Service definition document

  PDF

• Terms and conditions

  PDF

Request an accessible format

If you use assistive technology (such as a screen reader) and need versions of these documents in a more accessible format, email the supplier at ebba@scaleoutsystems.com. Tell them what format you need. It will help if you say what assistive technology you use.

Framework

G-Cloud 15

Service ID

488372991269747

Contact

Ebba Kraemer
+46766450939
ebba@scaleoutsystems.com

About the service

• Service categories:
  • Application Development and Deployment
    • AI platforms
      • AI life cycle
        • MLOps and Foundation Model Ops Software
• Multi cloud support: Yes

Service scope

• Software add-on or extension: No
• Cloud deployment model: Hybrid cloud
• Service constraints: Single-tenant SaaS delivery – Each customer is provided with an isolated, supplier-managed control plane environment.; ; Connectivity requirements – The service requires internet connectivity to operate and for the supplier-managed platform to retrieve updates and container images from trusted registries.
• System requirements: Client: Python 3.11+, internet connection

User support

• Email or online ticketing support: Yes
• Support response times: Scaleout responds to support questions during business hours (09:00–17:00 UK time, Monday to Friday, excluding public holidays). Initial response times depend on incident priority: critical incidents are acknowledged within 4 hours, high-priority incidents within 8 hours, normal issues within 2 business days, and general enquiries within 5 business days. Support requests received outside business hours or at weekends are logged and handled on the next business day, unless they relate to a critical control-plane outage.
• User can manage status and priority of support tickets: No
• Phone support: No
• Web chat support: Yes
• Web chat support availability: 9 to 5 (UK time), Monday to Friday
• AI chatbot: No
• Web chat support accessibility standard: None or don’t know
• How the web chat support is accessible: Web chat support is provided through dedicated, private messaging channels using web-based collaboration platforms such as Slack, Microsoft Teams, or Discord. These tools are accessible via standard web browsers and desktop clients and support text-based communication. Users can raise support questions, share logs or screenshots, and receive real-time responses from support engineers during business hours. For users unable to access web chat, support requests can also be submitted via email.
• Web chat accessibility testing: No formal accessibility testing with assistive technology users has been conducted specifically for the web chat channels. The platforms used provide baseline accessibility features such as keyboard navigation and screen reader support. We regularly gather user feedback on the effectiveness of our support channels and use this feedback to improve usability and accessibility where possible.
• Onsite support: No
• Support levels: Scaleout Edge Platform is provided with a single Enterprise support level, which is included in the subscription price. Support is available during UK business hours (09:00–17:00, Monday to Friday, excluding public holidays) via designated email and private web-based chat channels.; ; Support covers assistance with service usage, configuration within the platform, incident investigation, bug fixes, and guidance on operating the service effectively. Requests are prioritised by Scaleout based on impact and severity to the service.; ; Each customer is assigned a Technical Account Manager (TAM) who acts as the primary point of contact. The TAM provides proactive support, including onboarding guidance, architectural reviews, regular service health check-ins, and coordination of support activities.; ; Where required, customers have direct access to Scaleout support engineers and platform developers through the supported channels. There are no additional charges for different support levels; all Enterprise support services are included in the standard SaaS subscription.
• Support available to third parties: Yes

Onboarding and offboarding

• Getting started: Scaleout provides a dedicated onboarding manager to assist with initial project configuration and edge node registration. Scaleout provides a range of onboarding resources including user documentation and getting-started tutorials (html, pdf, video), online training, and onsite training. Standard onboarding typically takes 5–10 business days.
• Service documentation: Yes
• Documentation formats:
  • HTML
  • PDF
• Documentation accessibility standard: WCAG 2.2 AA
• End-of-contract data extraction: Self-Service Export. For a period of thirty (30) days following termination ("Export Period"), Customer shall maintain limited access to the Control Plane solely for the purpose of exporting Customer Content. Customer may export data using the following standard methods:; Models: Downloadable in original binary format via the WebUI, API and CLI.; Metadata & Metrics: Exportable in JSON format via the API and CLI.; 9.2 Bulk Export Services. Upon written request and subject to additional Professional Service fees, Scaleout can provide a full archive of the Customer’s environment, including:; A full export of the dedicated database.; A complete snapshot of the Object Store.; Data can be provided via secure Cloud Storage transfer (e.g., S3/Azure Blob) or, if requested for security/size reasons, via encrypted physical storage media at the Customer's expense.
• End-of-contract process: Upon notice of termination, a transition period of 30 days begins. During this time, the Service remains fully operational to allow the Customer to conclude active projects and prepare for data extraction.; The Customer retains full ownership of all trained models, training scripts, and metadata. These can be exported self-service via the API in standard, non-proprietary formats: JSON, CSV for metadata, and native binary representation for machine learning models. A bulk export of the database and the object store is also available at an additional professional services fee. ; Following the formal contract end date, Scaleout maintains the Managed Deployment in a "read-only" state for a 30-day grace period. ; Once the grace period expires, Scaleout initiates the final decommissioning of the single-tenant environment. This involves: ; Logical Deletion: All compute instances, storage volumes, and database records are deleted via the IaaS provider’s API, triggering the IaaS provider’s automated sanitization protocols (aligned with NIST 800-88 guidelines).; System and security audit logs are retained in Write-Once Read-Many (WORM) state for twelve (12) months post-termination.; A final invoice will be issued covering any outstanding charges or professional services. Once settled, a formal Certificate of Data Destruction can be provided to the Customer upon request.

Using the service

• Web browser interface: Yes
• Supported browsers:
  • Microsoft Edge
  • Firefox
  • Chrome
  • Safari
• Application to install: No
• Designed for use on mobile devices: Yes
• Differences between the mobile and desktop service: Mobile vs Desktop Service Differences:; ; The core Edge AI service is identical across platforms - both mobile (Android) and desktop (Python/C++) clients connect to the same backend infrastructure via gRPC protocol. Key differences are client-side only:; ; Mobile (Android Kotlin):; ; Native Android app integration; Optimized for mobile device constraints ; All on-device training occurs on customer-owned edge devices connecting to the supplier-hosted SaaS control plane; Desktop (Python/C++):; ; Full CLI and programmatic API access; Web dashboard access for orchestration and monitoring; Typically more computational resources available; Support for larger models and datasets; ; All clients use the same authentication (JWT tokens), gRPC communication protocol.
• Service interface: Yes
• Description of service interface: The platform provides five interfaces for interaction: (1) Web UI - Next.js dashboard with visual management of Edge AI workflows, real-time charts, and session monitoring accessible via browser; (2) REST API - RESTful endpoints (/api/v1/*) for programmatic access using JSON; (3) GraphQL API - flexible query/mutation endpoint for complex data operations; (4) gRPC API - binary protocol for high-performance edge client communicatio; (5) CLI - Python command-line tool (scaleout) for terminal-based operations. All interfaces share OAuth2/JWT authentication and provide access to models, sessions, clients, packages, metrics, and validation results. Users can choose their preferred interface based on use case.
• Accessibility standards: WCAG 2.2 A
• Accessibility testing: No formal interface testing with assistive technology users has been conducted to date. The web interface is built using Bootstrap 5 framework which provides baseline accessibility features including keyboard navigation and semantic HTML. Basic ARIA attributes are implemented for modals and interactive elements. However, we acknowledge that comprehensive testing with screen readers (JAWS, NVDA, VoiceOver), keyboard-only users, and other assistive technologies has not been performed. We recognize this as an important area for improvement and are committed to conducting formal accessibility testing and incorporating user feedback from individuals with disabilities in future development cycles
• User support accessibility: None or don’t know
• API: Yes
• What users can and can't do using the API: API Usage Description; Setup:; Users obtain OAuth2 access tokens via the Hydra OAuth2 server using client credentials grant with role:admin or role:client scope. Tokens authenticate all REST API and gRPC requests via JWT Bearer authorization headers.; Available Operations:; REST API (v1): Manage sessions, models, clients, packages, predictions, validations, and metrics through /api/v1/* endpoints; gRPC API: Real-time communication for edge clients (model uploads/downloads, status reporting, Task requests); GraphQL: Query and mutate data via /api/v1/graphql endpoint; Users Can:; Create/start training sessions and monitor progress; Upload/download machine learning models; Register compute clients and manage configurations; Query metrics, validations, and training statistics; Control training workflows (start/stop/pause); Limitations:; No runtime configuration changes: OAuth providers, database connections, and core settings require container restart; Authentication required: All endpoints except /health require valid JWT tokens in production; Role-based access: Admin role required for most operations; client role limited to edge client participation; No user self-service: User management handled through Kratos UI, not programmatically via API
• API documentation: Yes
• API documentation formats:
  • HTML
  • PDF
• API sandbox or test environment: Yes
• Customisation available: Yes
• Description of customisation: Customers can customize machine learning training scripts, machine learning model architectures, and federated learning aggregation logic. Modification of the underlying Control Plane infrastructure or OS is not permitted in the Managed Service.; ; Edge Learning Configuration:; ; Who: Project administrators via web UI, API, or CLI; What: Aggregation algorithms (FedAvg, FedAdam, FedYogi, FedAdaGrad), round timeouts, client participation requirements, validation settings; How: Session creation interface or API calls; ; Custom Training Code:; ; Who: Data scientists and ML engineers; What: Framework-agnostic compute packages (PyTorch, TensorFlow, Scikit-learn, custom frameworks); How: Package upload via web UI or CLI (scaleout package upload)

Scaling

• Independence of resources: The service supports independent scaling of resources within the supplier-managed SaaS environment. This includes supplier-managed load balancing, horizontal auto-scaling, resource policies, request rate-limiting, and alerting based on resource metrics. Scaling and resilience are handled by the supplier using underlying cloud platform capabilities

Analytics

• Service usage metrics: Yes
• Metrics types: Centralised, server-side infrastructure metrics are collected within the supplier-managed SaaS environment using industry-standard monitoring tooling (for example Prometheus and Grafana) and are presented via dashboards and APIs; ; Application metrics are provided in the form of performance metrics and logging directly in the software. Performance metrics are use-case and customer specific metrics, including various states of artifacts (performance of AI models), resource consumption on customer-owned edge devices connected to the supplier-hosted SaaS control plane, and task execution metrics.These metrics are available as both APIs and UI/Real-time dashboard. These metrics are designed for use-case/project specific users, such as ML engineers.
• Reporting types:
  • API access
  • Real-time dashboards
  • Reports on request
• Resource tagging: Yes
• FOCUS resource tagging: No

Supplier type

• Supplier type: Not a reseller

Staff security

• Staff security clearance: Other security clearance
• Government security clearance: Security Clearance (SC)

Asset protection

• Knowledge of data storage and processing locations: Yes
• Data storage and processing locations: European Economic Area (EEA)
• User control over data storage and processing locations: Yes
• Datacentre security standards: Managed by a third party
• Penetration testing frequency: At least once a year
• Penetration testing approach: Another external penetration testing organisation
• Protecting data at rest: Encryption of all physical media
• Data sanitisation process: Yes
• Data sanitisation type:
  • Deleted data can’t be directly accessed / Cryptographic Erasure
  • Data Erasure
• Equipment disposal approach: Complying with a recognised standard, for example CSA CCM v4.0, CAS (Sanitisation) or ISO/IEC 27001

Data importing and exporting

• Data export approach: User data (machine learning models, code bundles, artifacts, metrics, metadata) can be exported self-service using the webUI, API and CLI. Upon a support request, raw data export bundles can be delivered via secure cloud upload or encrypted physical media, if required; ; Data Export Methods:; Users can export data through multiple interfaces:; ; Models: Download trained models via web UI download button or REST API endpoint . Models are retrieved as binary files.; ; Metrics & Results: Access training/validation metrics through REST API or CLI returning JSON data. Export programmatically using Python API client ; ; Packages: Download compute packages via API
• Data export formats:
  • CSV
  • Other
• Other data export formats:
  • Machine learning models are exported in their native, binary formats.
  • Metrics and metadata can be exported as json.
• Data import formats: Other
• Other data import formats:
  • JSON
  • SavedModel/H5
  • TorchScript
  • ONNX

Data-in-transit protection

• Data protection between buyer and supplier networks:
  • Private network or public sector network
  • TLS (version 1.2 or above)
  • IPsec or TLS VPN gateway
  • Other
• Other protection between networks: Integrity by checksum
• Data protection within supplier network:
  • IPsec or TLS VPN gateway
  • Other
• Other protection within supplier network: Network policies, RBAC

Availability and resilience

• Guaranteed availability: Availability Target. Scaleout will use commercially reasonable efforts to make the Control Plane available with an Uptime Objective of 99.5% during each calendar month. This is a performance target and does not constitute a formal Service Level Agreement or a guarantee of service.; Definition of Uptime. For the purposes of this Schedule, "Uptime" is defined as the period during which the Scaleout Control Plane is accessible via the public internet. Uptime measurements apply strictly to the Control Plane and do not include the availability of individual Edge Nodes or local customer environments.
• Approach to resilience: Our managed deployments leverage Kubernetes as provided by the underlying IaaS provider for automated resilience:; ; - Self-healing: Automatic pod restart on failure with health checks and liveness probes.; - High Availability: Multi-replica deployments for critical services (API, controller, combiner, frontend).; - Rolling updates: Zero-downtime deployments with gradual rollout and automatic rollback on failure. ; - Resource management: Horizontal Pod Autoscaling adjusting to demand, preventing resource exhaustion. ; - Node failure recovery: Automatic pod rescheduling to healthy nodes.; ; Architecture Resilience:; ; - Stateless services: Frontend and API servers can scale horizontally without session affinity issues.; - Queue-based processing: Asynchronous task handling prevents cascading failures; - Graceful degradation: Services continue operating with reduced functionality during partial outages; ; Data Resilience:; ; - Regular automated backups with point-in-time recovery capability.; - Distributed object storage with multi-zone redundancy.; ; Datacenter Setup:; ; Managed deployments utilize multi-zone cloud infrastructure (GCP/AWS/Azure) with geographic redundancy options. Detailed datacenter specifications including SLA commitments available on request.
• Outage reporting: Scaleout employs automatic monitoring with email and chat alerts sent to authorized administrators when outages are detected. If an outage is detected, we begin investigation within 1h. If the outage is determined to be non-temporary, affected Users are notified with an email including an estimated timeline for resolution.

Identity and authentication

• User authentication needed: Yes
• User authentication:
  • Identity federation with existing provider (for example Google Apps)
  • Username or password
• Access restrictions in management interfaces and support channels: Access to management interfaces is restricted to authorised personnel only. Administrative access is granted on a named-user, least-privilege basis to designated supplier staff and approved customer administrators. Access requires authenticated credentials and is protected using strong authentication controls, network restrictions, and audit logging. All administrative access is logged and reviewed periodically.; Support channels are restricted to authorised users through invite-only, customer-specific channels to prevent unauthorised or cross-customer access. End users access the service through application interfaces only and do not have access to management or infrastructure controls.
• Access restriction testing frequency: At least once a year
• Management access authentication:
  • Identity federation with existing provider (for example Google Apps)
  • Dedicated link (for example VPN)
  • Username or password

Audit information for users

• Access to user activity audit information: Users contact the support team to get audit information
• How long user audit data is stored for: User-defined
• Access to supplier activity audit information: Users contact the support team to get audit information
• How long supplier audit data is stored for: User-defined
• How long system logs are stored for: User-defined

Security governance

• Named board-level person responsible for service security: Yes
• Software Security Code of Practice: Yes
• Security governance certified: No
• Security governance approach: Scaleout’s security governance is led by the CISO who reports directly to the CEO. While not yet certified, we seek to comply with ISO 27001 standards (and we plan to become certified in Q3 2026). We utilize a Shared Responsibility Model, leveraging the certified physical security of Tier-1 IaaS providers (AWS/Azure/Safespring). Internal oversight includes regular threat modeling, automated vulnerability scanning, and strict Least Privilege access controls. Infrastructure is managed via version-controlled templates to prevent configuration drift.
• Information security policies and processes: We are complying with ISO27001 (not yet certified).

Operational security

• Configuration and change management standard: Supplier-defined controls
• Configuration and change management approach: All service components are version-controlled in Git with immutable container images tagged and stored in container registries, providing full traceability throughout their lifecycle. Infrastructure configurations are managed as code using Terraform with versioned state tracking all deployment changes. All code changes undergo automated security scanning (Trivy) in CI/CD pipelines before deployment, identifying vulnerabilities and security issues. Pull requests require code review with automated testing including integration tests, security checks, and dependency vulnerability scans. Changes are deployed through staged environments with automated rollback capabilities. Security impact assessment is performed through static analysis, dependency checking, and automated security workflows that gate deployments.
• Vulnerability management type: Supplier-defined controls
• Vulnerability management approach: We perform code and container image vulnerability scanning using industry standard tools including GitHub. We are monitoring CVEs as they are published and we run regular risk management meetings. Critical patches are prioritized and deployed to the Managed Deployment within 10 business days of a verified stable release.
• Protective monitoring type: Supplier-defined controls
• Protective monitoring approach: We are monitoring with dedicated threat detection software. If a compromise is identified, we are following a documented process that all core developers are trained on. We respond to critical incidents within 1h.
• Incident management type: Complies with a recognised standard, for example, CSA CCM v4.0 or ISO/IEC 27035:2011 or SSAE-18 / ISAE 3402
• Post-quantum cryptography secure: No
• Incident management approach: We have a pre-defined process for common events that all key personnel are trained on. Users report incidents directly to the CISO, or the CEO.

Secure development

• Approach to secure software development best practice: Conforms to a recognised standard, but self-assessed

Public sector networks

• Connection to public sector networks: No

Pricing

• Discount for educational organisations: No
• Free trial available: No

Discount percentage by annual call-off contract value (excluding VAT)

• Less than £250,000: 0%
• Between £250,000 and £500,000: 5%
• Between £500,001 and £1,000,000: 10%
• Between £1,000,001 and £2,500,000: 15%
• Between £2,500,001 and £5,000,000: 20%
• Over £5,000,001: 25%

Standards and certifications

• ISO/IEC 27001 certification: No
• ISO 28000:2022 certification: No
• ISO 9001 certification: No
• Quality management systems (QMS): No
• CSA STAR certification: No
• PCI certification: No
• Cyber essentials: No
• Cyber Essentials Alternative: None of the criteria
• Cyber essentials plus: No
• Cyber Essentials Alternative: None of the criteria
• Other security certifications: No

Social value

• Mission: Kick start economic growth:

  To secure the highest sustained growth in the G7 - with good jobs and productivity growth in every part of the country making everyone, not just a few, better off.

  Policy Outcome 1: Fair work: That offers fair wages and good working conditions. Help people get a job, stay in work, and progress in their careers, with good employment opportunities across the country. Notes: Where there is a large proportion of labour costs in the contract, commercial teams should consider MAC in Outcome 1 as a priority

  • New or retained jobs on the contract workforce in the relevant area that meet the criteria set out in MAC 1b, 1c and 1d
  • Ensuring new workers are informed of their right to join a trade union
  • Activities that support an environment where staff are educated about, and feel empowered to, address their physical wellbeing
  • Activities that support an environment where staff are educated about and feel empowered to report and address bullying and harassment
  • Payment of more than the National Minimum Wage or National Living Wage (as appropriate) to the contract workforce
  • Monitoring and reporting of gender and ethnicity pay gaps and plans to address them where necessary
  • Support for educational attainment relevant to the contract, including training schemes that address skills gaps and result in recognised qualifications
  • Understanding of the modern slavery risks and issues affecting the market, industry, sector or country (of origin or of source) relevant to the contract, and the workforce in the supplier’s own organisation and those of its key subcontractors

  Policy Outcome 2: Skills for growth: Supporting growth sectors and addressing skills gaps.

  • Understanding of employment and relevant skills issues, and of the education and training issues relating to the contract. Illustrative examples: demographics, skills shortages, new opportunities in high growth sectors, geographic/local community and skills/employment challenges

  Policy Outcome 3: Resilient, innovative and flexible supply chains: Support economic growth through enabling resilient businesses, opportunities for small businesses and voluntary, community and social enterprises

  • Understanding of the types of businesses in the market and the level of participation by new businesses, entrepreneurs, start-ups, SMEs, VCSEs and mutuals
  • Understanding of local demographics, needs and opportunities for the co-design of the goods, services and works to be delivered under the contract
• Mission: Make Britain a clean energy superpower:

  To cut bills, create jobs and deliver security with cheaper, zero-carbon electricity by 2030, accelerating to net zero

  Policy Outcome 4: Sustainable procurement practices: Reducing carbon footprints, minimising waste, and promoting the use of clean energy and green technologies.

  • Collaborative ways of working with the supply chain to deliver additional environmental benefits in the performance of the contract, including working towards net zero carbon emissions and use of clean energy and green technologies
  • Understanding of how to influence staff, suppliers, customers, communities and/or any other appropriate stakeholders through the delivery of the contract to support climate and nature protection and improvement
• Mission: Break down barriers to opportunity:

  By reforming our childcare and education systems, to make sure there is no class ceiling on the ambitions of young people in Britain

  Policy Outcome 6: Employment and training: For those who face barriers to employment

  • Understanding of employment and skills issues, and of the skills and employment shortages of high growth sectors relating to the contract
  • Understanding of the issues affecting the representation of disabled people in the workforce in the market, industry or sector relevant to the contract, and in the supplier’s own organisation and those of its key sub-contractors
  • Working conditions which promote an inclusive working environment and promote retention and progression
  • Understanding of the issues affecting inequality in employment, skills and pay in the market, industry or sector relevant to the contract, and in the supplier’s own organisation and those of its key sub-contractors. Measures to tackle inequality in employment, skills and pay in the contract workforce

  Policy Outcome 7: Creating a pipeline of opportunities: For the contract workforce, reducing barriers to entry for under-represented groups.

  • Understanding of the issues affecting the development of new skills by target cohort
  • Advertising, promotional and outreach activities designed to raise awareness of the offer to reach the target cohort
  • Understanding of issues relating to entering the contract workforce
• Mission: Build an NHS fit for the future:

  That is there when people need it; with fewer lives lost to the biggest killers; in a fairer Britain, where everyone lives well for longer.

  Policy Outcome 8: Increasing productivity through physical and mental wellbeing: In the supply chain and communities in the relevant area.

  • Understanding of issues relating to health and wellbeing, including physical and mental health, in the contract workforce
  • Actions to invest in the physical and mental health and wellbeing of the contract workforce

Service documents

• Pricing document

  PDF

• Service definition document

  PDF

• Terms and conditions

  PDF

Request an accessible format

If you use assistive technology (such as a screen reader) and need versions of these documents in a more accessible format, email the supplier at ebba@scaleoutsystems.com. Tell them what format you need. It will help if you say what assistive technology you use.