OBSERVE
Understand infrastructure, applications, networks and system activity.
Glass Industries engineers advanced intelligence, security and computing systems for a world increasingly dependent on them.
FIG. 01 — GLASS SYSTEM ARCHITECTURECURRENT / 2026
INTRO / 001
Glass Industries is an independent deep-technology company building the intelligence layer for the systems everything else depends on — security, computing and infrastructure. We are early-stage. We work in the open where we can, and we do not claim capability we have not built.
Current focus left — exploratory scope right
SYSTEM / SENTINEL
AI for the systems that matter.
Glass Sentinel is a privacy-first intelligence platform designed to understand infrastructure, identify anomalies, investigate incidents and help operators respond. It is designed as an operational system — working alongside the people who run the systems that matter.
CAPABILITIES / 04
Sentinel is designed as a system that acts on infrastructure — not a conversation.
Understand infrastructure, applications, networks and system activity.
Correlate events and construct evidence-based explanations.
Use AI agents to analyse anomalies and determine likely causes.
Recommend or, with appropriate authorisation, execute corrective actions.
INCIDENT / CASE STUDY
Simulated scenario — how Sentinel would present an incident end-to-end. Not live production data.
Possible credential compromise.
Evidence correlated across
Temporarily revoke the active session and investigate the associated credentials.
PRIVACY / PRINCIPLE
Security products must be trustworthy before they are useful. Sentinel is being designed so that intelligence is gained through discipline, not through surveillance — the operator keeps the data, the logs and the authority.
These are design directions, not promises of perfection. Security is engineered, tested and earned over time.
Deployment designed to run where the infrastructure runs — inside the operator's boundary, not ours.
Telemetry protected in transit and at rest. Data minimised and retained on the shortest sensible schedule.
Actions are logged, evidence is traceable and decisions carry attribution that can be examined.
Sensitive actions keep a human in the loop. The system recommends; the operator authorises.
PRINCIPLES / 001
Advanced technology should be demonstrated through working systems, rigorous engineering and measurable results — not declarations. This is the standard Glass Industries holds itself to, internally and publicly.
gl / operating-standard 1.0
Turn difficult ideas into functioning systems. Working software and hardware are the primary way we communicate.
Test assumptions, measure performance and expose weaknesses. Claims are only as valuable as the evidence behind them.
Take successful systems from prototype to dependable infrastructure — carefully, deliberately, and never by declaring it done.
VISION / LONG TERM
Glass Industries is building toward a broad deep-technology company. Most of the areas below are future exploration — only what we have actually built is presented as current.
Privacy-first AI for infrastructure intelligence, anomaly detection and incident investigation.
The model and agent infrastructure powering Sentinel today; a platform being built in service of real systems.
Distributed compute and embedded systems that sit where the infrastructure is, not where it is convenient.
Autonomous systems for physical environments. Early research direction, not a product.
Sensing and communications for constrained, high-integrity environments. Long-horizon research interest.
The laboratory function — publications, benchmarks and experimental systems feeding every line of the company.
Status key — active / in development = current work · exploring = direction · research = investigation
gl / roadmap 0.1RESEARCH / LAB
An engineering organisation that publishes technical work. Research here feeds directly into the systems we build — it is not decorative.
Publication platform in preparation — results and methods will be shared as they mature.
COMPANY / ABOUT
Glass Industries is an independent deep-technology company focused on building advanced systems across artificial intelligence, security, computing and emerging technologies. We are early-stage and we intend to stay close to the engineering — the company is being built the same way its systems are: first principles, working prototypes, then dependable infrastructure.
If this site ever implies otherwise, that is an error. We would rather be underestimated early than overstate late.
CONTACT / DIRECT
Serious problems, long horizons. If you engineer infrastructure, operate systems, fund deep technology or research the same problems we do — we want to hear from you.