Applied AI & security engineering
We build systems that think — and prove they hold.
Lilly Systems researches, develops and delivers intelligent software and physical AI systems. We also test and harden the infrastructure they run on. Both halves of that work sit with one team, which is why the second one is not a surprise at the end of the first.
- Production systems, not prototypes
- Findings you can re-verify
- Scoped and authorised testing
- EN · DE · FR · IT
Lilly Schweizer Systems LLC · Albuquerque, New Mexico
What we do
Five disciplines, one engineering standard.
Two of them build something new. Three of them make sure that what exists can take a hit. They are listed in the order they usually happen, though most engagements start somewhere in the middle.
Intelligent Software
Applied AI that runs in production: retrieval and reasoning pipelines, decision support, and process automation, integrated with the systems you already operate.
Read morePhysical AI Systems
Perception, sensor fusion, on-device inference and control loops for machines and edge hardware — with the failure behaviour worked out before deployment, not after.
Read moreCybersecurity Audits
A structured review of architecture, code, configuration and process against recognised baselines, ending in a prioritised report every finding of which can be re-checked.
Read moreSecurity Assessments
Authorised, scoped technical testing of applications, APIs, cloud and networks: threat modelling, penetration testing and red-team exercises, rated by real business impact.
Read moreInfrastructure Hardening
Turning findings into an enforced baseline: least-privilege identity, network segmentation, secrets management, patch and backup discipline, logging and detection — as code.
Read moreWhy both
Building and securing are the same skill, applied twice.
A team that has never had to defend a system builds different software than a team that has. We do both on purpose: the audit work keeps the engineering honest, and the engineering work keeps the audits grounded in what is actually buildable.
Build
Systems that make decisions and act on them — in software and in hardware.
- Applied AI in productionRetrieval, reasoning and automation wired into the systems you already run.
- Perception and controlSensor fusion, on-device inference and control loops with defined failure behaviour.
- Owned, not rentedYou get the source, the deployment path and the documentation to run it without us.
Secure
Evidence about how a system actually behaves — and the work to change that behaviour.
- Audits against a baselineArchitecture, code, configuration and process, measured against a standard rather than a hunch.
- Testing under authorisationThreat modelling and penetration testing inside a written, agreed scope.
- Hardening that persistsBaselines expressed as code, so the state is enforced instead of documented.
How it runs
Research, develop, deliver, operate.
The same four stages whether we are building a system or breaking one. What changes is which stage carries the weight.
Research
Understand the domain, the constraints and the failure modes before proposing anything. This is where scope gets honest.
Develop
Build in short, reviewable increments. Every increment is something you can run, not a status report.
Deliver
Hand over source, infrastructure definitions and documentation. Delivery means you can operate it without us.
Operate
Support, monitoring and periodic re-testing — optional, and never a condition of getting your system.
- Legal entity
- Lilly Schweizer Systems LLCTrading as Lilly Systems
- Registered office
- Albuquerque, New MexicoUnited States
- Working languages
- EN · DE · FR · ITReports and workshops in any of the four
- Engagement models
- Project, retainer, assessmentScope agreed in writing before work starts
Let us look at your system.
Tell us what you are building or what you need reviewed. You get an honest assessment of scope and effort before anyone signs anything.