Glossary | LLM Machines

AI & Models.

The model layer: what is being run, how it represents text, and the tuning vocabulary around it.

LLM: Large Language Model. A neural network trained on massive text corpora to predict the next token in a sequence. Today's chat assistants are LLMs running with instruction tuning and safety layers on top.
Open-weight model: A model whose trained parameters are publicly released, so the model can be downloaded, audited, fine-tuned and run locally. Examples: Llama, Mistral, Qwen, DeepSeek. Contrast with proprietary models accessible only through a vendor's API.
Token tokenization: The atomic unit a language model reads and writes. One token is roughly ¾ of an English word. Tokenization is the preprocessing step that splits text into these units.
Context window: The maximum number of tokens a model can consider at once when generating a response. Larger context windows let the model reason over longer documents but raise memory and latency costs.
Inference: Running a trained model to produce output, as opposed to training, which produces the model in the first place. On-prem inference means the model is executed on your own hardware.
Fine-tuning: Adapting a pre-trained model to specific tasks or domain language by continuing training on a smaller, targeted dataset. Distinct from prompt engineering, which shapes input rather than weights.
LoRA / QLoRA: Low-Rank Adaptation. A parameter-efficient fine-tuning method that trains a small set of additional weights instead of updating the full model, drastically cutting compute and storage. QLoRA adds quantisation for further savings.
Quantization: Reducing the numerical precision of model weights (for example, from 16-bit to 4-bit) to shrink model size and accelerate inference, usually with a small accuracy trade-off.
Hallucination: When a model generates plausible-sounding but factually incorrect output. Production systems address this with source verification, careful prompting and human review.
Temperature: A sampling parameter that controls randomness in model output. Zero produces deterministic, focused responses; higher values produce more creative or varied output.
Llama · Mistral · Qwen · DeepSeek: Major families of open-weight LLMs from Meta, Mistral AI, Alibaba and DeepSeek respectively. The four that most enterprises evaluate first when planning local inference.

Architecture & Infrastructure.

How an AI system is assembled, served, and bounded, from the appliance level down to individual protocols.

On-prem on-premise: Software or hardware that runs inside infrastructure physically controlled by an organisation, rather than in a third-party public cloud. The architectural opposite of SaaS.
Appliance: A pre-integrated hardware-plus-software unit shipped as a single product. Network firewalls, storage arrays, and our AI platform are all appliances; the customer does not assemble the components.
Managed inference core: The primary supported service boundary for running approved local models. LLM Machines configures and maintains model serving, the API gateway and policy controls, the operational Console, the observability layer, and lifecycle and recovery tooling inside the Customer's infrastructure.
API gateway: The local service between Applications and AI models, handling authentication, model allowlists, routing and usage metadata through the documented Application API.
Chat Completions: The documented API protocol your favourite applications and harnesses use to access approved local models. It supports streaming and non-streaming inference through /api/app-gateway/v1/chat/completions; available models are listed through /api/app-gateway/v1/models.
Customer application: Your preferred AI application or harness connected to the managed inference core through a dedicated credential and the documented Application API. You choose and operate the interface and workflow while LLM Machines provides the local inference contract.
Model alias: A stable model name exposed to an Application while the Core Appliance maps it to the approved local model and inference route. This lets backend models change without changing the connected Application.
Support boundary T1 / T2: The managed inference core is the primary supported service. Customer applications, their data, permissions and workflows remain outside the standard Product and connect through the documented API contract. Optional integration consulting is separately scoped.
Emergency Isolation Mode: An administrative control that blocks Application inference traffic at the Product edge while keeping the Console, identity, audit, hardware monitoring, backup and recovery available.
Signed offline update: A manually imported software bundle whose signature and compatibility are verified before installation. Updates use preflight checks, a local snapshot, health validation and rollback without a cloud call-home.
Core Appliance: The supported customer-premises deployment: Product edge, managed inference core, operational Console, identity, observability, lifecycle and recovery controls inside Customer-controlled infrastructure.
Agent agentic workflow: An application-level pattern in which an AI system plans, decides and calls tools to complete a multi-step task. It is not a bundled or first-party LLM Machines Product surface.
Tool-call transport function calling: The Application API can carry structured tool-call messages produced by a model. Customer applications and separately scoped advanced integrations can use those messages to continue tool-enabled workflows.
Source-available first-party Product source: Original first-party LLM Machines Product source licensed under the unmodified PolyForm Internal Use License 1.0.0 for customer internal business use and internal changes. It may not be distributed, sublicensed or transferred under that licence. Third-party software remains under its upstream licence.
Observability layer: The self-hosted monitoring, metrics and alerting layer used to understand appliance health, inference capacity and operational events without retaining workload content.
Self-hosted websearch: A locally controlled plugin that authorised Applications can use under the Customer's network policy, credentials and operational controls.

Privacy, Security & Identity.

Concepts that decide where data sits, who can reach it, and how the Core Appliance proves that nothing leaves your infrastructure.

PII: Personally Identifiable Information. Data that can identify an individual, such as a name, email address, national identifier or, in some cases, an IP address.
Pseudonymisation vs anonymisation: Pseudonymisation replaces identifiers with reversible tokens; it remains personal data under GDPR. Anonymisation removes identifiers irreversibly and the result is no longer personal data. Most "anonymisation" in industry is actually pseudonymisation.
Data residency: A requirement that data physically resides in a particular geographic location. It is often confused with sovereignty: residency is about where bytes sit, while sovereignty is about which legal regime governs them.
Data sovereignty: The principle that data remains subject to applicable jurisdiction, access rights and operational controls. Physical location matters, but it does not by itself eliminate foreign-government access risk.
Sovereign AI: AI systems where the model, the infrastructure it runs on, and the data flowing through it remain under the user's jurisdictional and operational control. The product category LLM Machines sells.
Air-gapped: A deployment with no network connection to external systems. Air-gapped AI runs purely on internal data and updates ship via removable media. Used in defence, intelligence and critical infrastructure.
Zero-trust architecture: A security model that verifies every request as if it came from an untrusted network, regardless of origin. The maxim: never trust, always verify.
RBAC: Role-Based Access Control. Permissions are assigned to roles (e.g. "engineer", "admin"); users inherit the permissions of their roles. The standard enterprise authorisation model.
SSO: Single Sign-On. A user authenticates once with a central identity provider and accesses many applications without re-entering credentials.
SAML · OIDC: SAML is the XML-based SSO protocol enterprises typically use with legacy identity providers. OIDC (OpenID Connect) is the modern JSON-based equivalent built on OAuth 2.0; preferred for new applications.
Identity federation: A trust relationship between systems that lets a user authenticated by one system be recognised by another, without sharing passwords. The mechanism behind SSO.
mTLS: Mutual TLS. Standard TLS proves the server's identity to the client; mTLS additionally proves the client's identity to the server, using certificates on both sides. Common between trusted internal services.
Application credential: A dedicated static key or OAuth client credential used by an Application to access approved model aliases through the Application API. Credentials can be rotated or revoked without changing the connected workflow.
Activity & Audit: The Console section for metadata-only operational and security events, with signed JSON and CSV exports. It records who changed which resource and when, without retaining prompts, model responses or tool content.
Operational metadata: Service data used to operate and support the appliance, such as timestamps, application identifiers, model aliases, token counts, latency, status, capacity signals, and administrative events. Workload content such as prompts, model responses, tool arguments, tool results, and request or response bodies is kept outside this metadata set.

Regulations & Compliance.

The EU framework that increasingly shapes which AI architectures are acceptable, plus the global standards enterprise buyers check first.

EU AI Act: The European Union's regulation classifying AI systems by risk and imposing obligations on developers and deployers, especially for high-risk uses. In force from 2024 with phased enforcement through 2026–2027. The first comprehensive horizontal AI law in any major jurisdiction.
GDPR: General Data Protection Regulation. The EU's comprehensive data-protection law, effective 2018. Establishes rights for data subjects and obligations for controllers and processors; fines can reach 4% of global annual turnover.
NIS2: Network and Information Security Directive 2. The EU's cybersecurity framework for essential and important entities. Adds board-level accountability and supply-chain security requirements. Must be transposed into national law by each member state.
EU Data Act: Regulation on harmonised rules for access to and use of data, in force from 2024. Targets cloud lock-in by mandating portability and switching rights for cloud services.
Schrems II: The 2020 Court of Justice of the EU decision that invalidated the EU-US Privacy Shield. Transatlantic transfers require a valid transfer mechanism and, where necessary, supplementary safeguards based on the specific processing context.
DPA: Data Processing Agreement. The contract between a data controller and a data processor mandated by GDPR Article 28. Note: DPA also stands for Data Protection Authority, the national regulator. Context determines meaning.
DPIA: Data Protection Impact Assessment. A GDPR-mandated risk assessment performed before deploying systems that handle personal data at scale or with elevated risk.
Data controller · Data processor: Under GDPR, the controller decides why and how personal data is processed; the processor acts on the controller's instructions. AI vendors are usually processors; their enterprise customers are usually controllers.
Right to be forgotten: A GDPR right (Article 17) allowing data subjects to require deletion of their personal data under specified conditions. Creates real engineering challenges for AI systems trained on personal data.
SOC 2: Service Organization Control 2. A US audit framework focused on five trust principles: security, availability, processing integrity, confidentiality and privacy. The common B2B SaaS compliance baseline.
ISO 27001: The international standard for information security management systems. Certification is frequently requested in enterprise and public-sector procurement, depending on the buyer and scope.

Operations & Deployment.

The vocabulary of getting an AI appliance from contract to live service, and the commercial concepts shaping the buy decision.

Discovery: The opening phase of an enterprise deployment, where we map models, connected Applications, performance targets, identity, network topology, hardware and approved network policy. It drives sizing and configuration decisions downstream.
Sizing: Calculating the hardware footprint (GPU count, RAM, storage) needed to meet your performance and concurrency targets.
Pre-flight check: Validating that all prerequisites, including network, identity, storage and power, are in place before the appliance ships.
Smoke test: A fast end-to-end test confirming a freshly installed system performs core functions correctly. It is a focused validation that catches broken deployments early.
Self-test: Built-in diagnostics that the appliance runs on boot to verify hardware health and software integrity.
Pre-shipment: Everything that happens before hardware leaves our integration facility: image preparation, signed release-manifest assembly and recovery-material preparation.
API handoff: The completion point for standard onboarding, when validated Models and Chat Completions endpoints, application credentials, model allowlists, operating documentation, and acceptance results are handed to the Customer for use by its chosen applications.
Support session: An explicitly authorised, time-limited access window used to diagnose or resolve an agreed issue. Its scope, approvals, logging, and revocation follow the signed support terms and the Customer's access controls.
Outcome-led consulting: A separate engagement organised around a defined business or integration outcome, a bounded delivery, measurable acceptance criteria, and documented handoff. The Customer selects and operates the accepted application, tools and workflow after handoff.
SLA: Service-Level Agreement. The contractual commitment to specific availability, response-time or performance targets.
Vendor lock-in: Dependence on a single vendor's platform such that switching becomes prohibitively expensive or technically infeasible. The EU Data Act explicitly targets cloud lock-in.
Pilot proof of concept: A bounded, time-limited deployment to validate the appliance against your real workloads before committing to a full rollout.

Glossary.

AI & Models.

Architecture & Infrastructure.

Privacy, Security & Identity.

Regulations & Compliance.

Operations & Deployment.

Want this in your stack?