ENG-012 weeks
Discovery sprint
A fixed-scope fortnight to rank use cases, audit your data, measure candidate models and size the hardware. It ends with a written plan, a configuration sheet and a quote for the build.
CSTM-PRC · REV AcstmAI™ · Process
The AI development process, start to finish.
Two weeks of discovery, a fixed-scope build measured against your own cases, a hardware burn-in, installation on your site, and a handover that leaves you owning all of it.
Fig. 01 · Lathe area of a machine shop, Paterson, NJ
Build process
From first call to running in your rack.
Each phase ends with a review and a go/no-go. A first system is typically in production 8–12 weeks after discovery begins.
012 weeks
Discovery
Pick the workflow, audit the data, size the hardware, write the evaluation set.
- Use-case ranking
- Data audit
- Configuration sheet
- Quote
024–7 weeks
Build
Software built and scored against the evaluation set, reviewed weekly with your team.
- Weekly demos
- Eval scores
- Integrations
03≈ 1 week
Burn-in
Hardware assembled and run under sustained load on our bench, with the software loaded.
- Stress tests
- Imaging
- Acceptance run
041–2 weeks
Install
Racked, cabled, networked and joined to your identity provider, on your site.
- Site install
- SSO + roles
- Handover
05Ongoing
Operate
Your team runs it with our runbooks, or we run it for you. Models updated as they improve.
- Monitoring
- Model updates
- Quarterly evals
Typical first system: 8–12 weeks to productionThe full process →
How we work
Engagement models.
Pick the shape that fits your team. Most projects use the first two, and some add the last two.
ENG-02Per phase
Fixed-scope build
Agreed deliverables, built against an evaluation set and demonstrated weekly. Scope changes are written down and re-quoted, not absorbed silently.
ENG-03Flexible
Co-build with your team
Our engineers work alongside yours in your repositories, so your team learns the stack while it ships, and owns it afterward.
ENG-04Ongoing
Managed operations
After launch we monitor, update models, re-run evaluations and handle hardware issues, for teams who would rather not staff it.
Process
Three rules we don't bend.
PRC-01
Evaluation-first delivery
Before we write the system, we write the test: real examples from your business with the answers your experts would accept. Every model, prompt and change is scored against it, and the scores decide what ships. You see the numbers every week.
PRC-02
Security review readiness
We assume your security team will review the system, so we prepare for it from the first week: an architecture diagram, data-flow description, access model, logging and retention settings, and a list of every permission the system holds.
PRC-03
Handover
You own what we build: the code, prompts, evaluation sets and any fine-tuned weights, delivered to your repositories and storage, plus runbooks and training for the people who will run it.
Reel 02 · Milling and boring area, Paterson, NJ1994
FAQ
Process questions.
What does the discovery sprint produce?
A ranked list of use cases, a data audit, model test results on your examples, a hardware configuration sheet, a build plan and a quote. If the numbers don't support a build, the plan says so.
How often will we see progress?
Weekly: a working demo and the current evaluation scores. Nothing is held back for a big reveal.
Can we stop after a phase?
Yes. Each phase ends with a go/no-go, and you keep everything delivered up to that point.
Get a quote
Spec your system.
Tell us the models you want to run, how many people will use them and where the hardware should live. An engineer replies with a first configuration and the questions that decide the quote.