Prepared for Enterprise Operations Classified IT
Give programs a platform worth onboarding to.
One vertical slice at a time.
A 12–14 week thinnest viable platform, built with your tech leads and engineers. MATLAB goes first: from the test bed into air-gapped production on a STIG-hardened baseline, with segmented program environments. The next priority apps follow on the same pipeline, and to a cloud target, onboarded by your own team.
What we heard
Programs are told to onboard, but the platform doesn't have what they need yet
Company policy says every program onboards to the Enterprise Operations platforms. But the engineering apps programs ask for aren't there yet, so exceptions keep piling up. Meanwhile DoD initiatives are pushing isolated program networks to consolidate and connect to the DoD clouds.
The gap
Exceptions instead of onboarding
The mandate
On-prem, cloud native, declarative
The targets
One pipeline, many destinations
The separation
Special access programs stay discrete
The team
Lean, strong and stretched
The open debate
Working models over more design
Our approach
Start with the thinnest platform that gets one real user what they need
Instead of designing the whole platform up front, we agree on one clear user outcome and build only what it takes to deliver it. Each slice takes a priority app all the way to the people who use it, and adds only the platform pieces that app needs. Every slice adds to the reference architecture, the decision records and the security package.
Thinnest
Viable platform
Pull-based GitOps, declarative IaC, quality and security gates in the pipeline
Vertical
Slice by slice
Each priority app goes from dev through test into air-gapped production
Owned
By your engineers
Paired every week, so your team runs and extends it after we leave
The platform path
- Pull-based GitOps on CAKE, with infrastructure as code from the first commit
- Quality gates and security controls built into the pipeline, on a STIG-hardened baseline
- Promotion from development through test into disconnected, air-gapped production
- Cloud native from the first slice, so the same pipeline reaches AWS and Azure DoD
Needed from Lockheed Martin
Platform and Kubernetes tech leads pairing weekly, plus a cyber architect to review controls as they're built.
Apps to the people who use them
- App priorities set with the director, the architects and the business area programs, each tied to a real user
- MATLAB and its modules first, then the next priority apps (Cameo, Teamwork Cloud, DOORS, Creo)
- Desktop delivery for engineering apps, choosing the virtual desktop with you (Guacamole, TGX, or others)
- A self-service front end where programs request and onboard apps, with the portal tool picked to meet your compliance needs
Needed from Lockheed Martin
The app priority list, and access to the first MATLAB users in a business area program.
Separation, security and AI
- Segmented program environments on the central platform that meet special access program separation rules
- A second program environment in production to prove the separation works
- A security assessment package built up with every slice, to support accreditation
- AI-assisted test and quality-gate building, using only Lockheed Martin's approved AI systems
Needed from Lockheed Martin
Cyber architects to accept the security package, and an introduction to the Lockheed Martin AI Center.
One vertical slice
Every app travels the same path
Step 1
Request
Program asks for the app through the front end
Step 2
Build
Declarative IaC + GitOps repo
Step 3
Gate
Quality, security and STIG checks
Step 4
Promote
Dev → test bed → air-gapped prod
Step 5
Use
Engineer launches it in their program's environment
12–14 weeks
Three milestones, each one real users can see
Week ranges are estimates and will be set in the SOW. Each milestone is reached when someone outside the project team can use the result, not when a document is finished.
Weeks 1–4
On the test bed
An engineer uses MATLAB delivered through the thinnest pipeline into one program environment. Constraints that need longer approvals are simulated. Every deliverable starts from this slice, and it sets the scope and success measures for taking MATLAB into production.
Done when that engineer has used MATLAB this way and their feedback has shaped what comes next.
Weeks 5–9
In production
MATLAB goes into air-gapped production on the STIG-hardened baseline, running disconnected. A second program environment proves the separation. Every deliverable is updated with what production taught us.
Done when program engineers are using MATLAB there and your cyber architects accept the security assessment package.
Weeks 10–12
A repeatable pattern
We package the MATLAB path into a documented, repeatable onboarding pattern your engineers can run, and show the same pipeline deploying to a cloud target. We start discovery on the next priority apps and outline the next phase of work to onboard them.
Done when the onboarding pattern is proven with MATLAB (or another representative app), handed over to your engineers, and the next phase is outlined with the director.
Outcomes and deliverables
What you'll have at the end
Outcomes
- Program teams can onboard and actually use the platform, because the engineering apps they need are there.
- The same pipeline deploys an app on-prem and to AWS or Azure.
- A priority app moves from development through test into air-gapped production on a STIG-hardened baseline.
- Enterprise Operations engineers onboard the next app themselves.
Deliverables, grown slice by slice
Operational proof of concept
Pull-based GitOps implementation
Validated reference architecture
Architecture decision records
Security assessment package
Multi-tenancy validation report
Operational runbooks
Financial analysis
Executive recommendations for funding
What we need from you
- Tech leads and engineers building with us every week
- An early start on access and background checks
- Access to the test bed and each environment a slice reaches
- The app priority list and the users behind it
- The director and architects in the design decisions
- An introduction to the group that runs your internal AI systems
Not included
- Ongoing managed services after the engagement (your team runs what goes into production)
- Formal accreditation of the platform (the security package supports it)
- Buying or licensing apps and tools
- Custom AI model development or hosting
Why Liatrio
We build it with your team, so the capability stays with your team
Working software from the first weeks
An engineer is using MATLAB on the test bed in the first month. The architecture decisions and the security package come from what actually ran, not from slides.
Continuity, not dependency
We pair with your engineers, demo continuously and hand over step by step. That speaks directly to the director's continuity concern and five years of upskilling effort. The finish line is your team onboarding an app without us.
Platform engineering for regulated enterprises
Platform as a product
Built for its users, the programs, with self-service onboarding as the goal.
Thinnest viable platform
Only what the current slice needs, so nothing gets over-built.
Compliance built in
STIG, quality and security gates and evidence produced by the pipeline itself.
AI within your rules
Using only your approved systems. Our Anthropic, Cursor and OpenAI partnerships are there if you want them later.
Next steps
Moving while this year's budget is available
Investment
$500K – $650K
A 12–14 week, outcome-focused engagement. The final price within this range will be set once the first slice, the target platform and the team size are agreed with Sean and the director.
- 1Put the NDA in place and receive the director's outcomes and deliverables list.
- 2Review this proposal with Sean, then prepare together for the director meeting.
- 3Meet the Lockheed Martin AI Center to confirm which approved AI systems we can use.
- 4Confirm the first MATLAB users, the platform for slice one, and start access requests.