Systems engineering for AI strategy
The model is the smallest part of an AI system. Systems engineering covers the rest: the data, the decisions, the people, and the payoff.
Walker Armistead · StatusNext
The math before the model
Bringing Systems Engineering to the decisions that are expensive to get wrong: AI strategy, contract economics, and large technology transformations.
sla_capability.sim
Achievable service level · 10,000 simulated months
- P(hit 97.5% ask)
- 0.05
- P(hit 93% commit)
- 0.89
- Expected penalties
- −84%
What I do
Economic decision analysis
AI is pushing service businesses toward fixed-price work. Like any fixed fee, service level, or performance contract, that work gets priced on the typical case, and the hard cases decide who absorbs the cost. Model the full range of outcomes before committing, and bring the math to the negotiation.
Building & integration
Data pipelines, simulations, decision models, and working prototypes are a de-risking strategy: each one retires a specific risk while it’s still cheap, before the full program is funded. Then hands-on support carries the program through integration, including the data migration.
Twenty years across aerospace and enterprise software.
More about me →Founder & principal, StatusNext LLC
Senior account executive and sales manager, Manufacturing, Enterprise Data & AI Leader
Analytics leader, Major DoD Program, Aerospace & Defense, Sustainment
Operations research analyst, Aerospace & Defense R&D
Adjunct professor, Manufacturing Engineering, Texas Christian University
F-16 crew chief, 187th Fighter Wing, Alabama Air National Guard
Selected work
Service contracts · Lockheed Martin
Showed leadership that the service levels on the table depended on the customer, not on us, before we signed.
Simulation showed the proposed metrics depended on customer behavior, not on our own performance. The business negotiated achievable terms and hit its financial targets on the contract.
Read more
Service contracts, round two · Lockheed MartinRenegotiated the same contract so the customer finally got what it wanted.
New metrics the end customer valued and the business fully controlled. A satisfied customer and a profitable contract. Leadership brought me back for this one, and again asked me to present the analysis at the table.
Read more
Fleet & network planning · Lockheed MartinRedesigned a cargo fleet's distribution network after a major aircraft upgrade.
The customer adopted the new hub-and-spoke model, realized the projected fuel savings, and avoided equipment purchases the old model would have required.
Read more
Data migration · StatusNext clientMoved a small independent school from disparate legacy systems to a single consolidated platform.
The school had no migration skills in house and no migration service from the new vendor. It switched over with no downtime, onboarded on time, and avoided hundreds of hours of manual data work.
Read more
What I'm building
Golf analytics
Pre-tournament finish probabilities (win, top 5, top 10, top 20, made cut) from a Monte Carlo simulation grounded in strokes gained.
- Python
- Flask
- React
- MySQL
- Airflow
- MLflow
- Monte Carlo simulation
CollegeCompass
College counseling app. Turn GPAs and test scores into confident, data-backed college recommendations for every student.
- Next.js
- Prisma
- Postgres (Neon)
- Clerk
- Stripe
- Resend
- Airflow
What will your promise cost?
An interactive model of fixed fees and service levels. Put in your own numbers and see how a price set on the typical case holds up across the full range of outcomes.
- TypeScript
- Monte Carlo simulation
- Astro
Let’s run the numbers before you commit.
A few sentences is plenty: what you’re weighing, when you need to decide, and what it costs if you get it wrong. I read and reply to every message myself.