Design 10x

From concept to a working prototype, in weeks.

Senior device design engineering with AI running through every step — and a working prototype in your hand at the end of it, not a render.

10X-SVC-DESIGN REV B
What
Design Sprint
Price
Fixed, quoted in 48 h after review
When
4 to 6 weeks

AI-drafted · evidence-cited · human-reviewed

Why this holds up

01

Senior development engineering, not junior throughput

The engineers on the sprint have run device development across catheter-based and Class II and Class III programs. The person doing the design work is the person you meet at the review.

02

In-house machining, so sprints end in hardware

Prototyping and machining are in-house. That is why a sprint ends with a part you can hold and test rather than a render you have to imagine, and why the turnaround does not depend on a supplier queue.

03

AI accelerates, a named human signs

AI widens the concept set, drafts the requirement and test-plan structure, and does the tedious cross-checking. Every deliverable carries a named human reviewer who signed it, and the review record says what they changed.

How it works

  1. DES-01

    Inputs and scoping review

    Your intended use, constraints, and whatever requirements exist. The review is where we agree the scope, quote it, and tell you if this is not a sprint we should take.

  2. DES-02

    AI-assisted concept set

    A deliberately wide set of concepts generated against the requirements, then narrowed by engineers on manufacturability, risk, and the pathway — with the reasoning written down.

  3. DES-03

    CAD and machined prototype

    The surviving concept goes to CAD and then to the machines. You see the design before it is cut, and you get the files as well as the part.

  4. DES-04

    Bench data and design review record

    A first bench characterisation against the requirements, and a design review record your quality system can file rather than reconstruct later.

What the work looks like

The design sprint, requirements through design review record Requirements enter an AI-assisted concept generation step, which produces a wide concept set. A named engineer review gate narrows that set to the surviving concept, with the reasoning recorded. The surviving concept goes to CAD, then to in-house machining as a prototype, then to bench characterisation against the requirements. A second named engineer review gate produces the design review record. Both review gates are human decisions, drawn at the same weight as the AI step. Requirements + intended use Concept set AI-widened AI WIDENS Engineer review what survives, and why HUMAN DECIDES CAD files are yours Prototype machined here Bench data vs requirements Design review named engineer signs HUMAN DECIDES Review record filed, not rebuilt bench data can move a requirement Fixed price, quoted 48 h after the scoping review · 4 to 6 weeks · you keep the IP
FIG. 1 The sprint, end to end. The human-review gates are drawn as prominently as the AI steps because that is their weight in the process: AI widens the set, people decide what survives.

The offer

The pilot engagement

This is a new service line and we are pricing it to learn: a fixed fee, quoted within 48 hours of the scoping review, for a single concept or subsystem.

Intake is limited. A sprint occupies a senior design engineer and machine time for four to six weeks, and neither parallelises, so we run a small number concurrently. If the concept is larger than one subsystem, or if it is not one we can do well, we say so at the review rather than after an SOW.

Three smaller engagements exist for teams that do not need a full sprint: a fixed-price DFM and design review, one to two weeks; a prototype sprint for engineering teams that have the design and need the parts; and a fractional senior design engineer for programs that need the seniority without the headcount. All three are quoted the same way, after a review.

What you receive

  • A machined prototype of the surviving concept, in a representative material, plus the CAD files.
  • The concept set that was considered and the written reasoning for what was eliminated — the part that usually lives in somebody’s head.
  • A requirements set and a test plan structured so verification can be scoped from it directly.
  • First bench data characterising the prototype against those requirements.
  • A design review record, and the human-review record below.

What the human-review record contains

  • The named engineer who reviewed each deliverable, and the date.
  • Every AI-generated concept or draft that was materially changed on review, and what changed.
  • The rationale for the concept that survived, and for the ones that did not.
  • Decisions escalated to you rather than made for you, and what you decided.
10X-SVC-DESIGN REV B
What
Design Sprint
Price
Fixed, quoted in 48 h after review
When
4 to 6 weeks

AI-drafted · evidence-cited · human-reviewed

Also inside the alliance

Capabilities we deliver that do not have their own page yet. No prices and no lead times here — those belong to an engagement, and an engagement starts with a review.

  • Engineering simulation

    FEA, CFD, and CT-based analysis, with the credibility evidence structured to ASME V&V 40 and to what FDA expects of computational modelling in a submission.

  • Machine-vision final test

    End-of-line inspection stations, with the vision models running on hardware inside your building, delivered with the TMV documentation that makes the station audit-ready.

  • Production automation & AI optimisation

    Automation roadmaps, line concepting, and data-first yield and throughput analysis built from your existing historian and MES exports.

  • Local / on-prem AI deployment

    Our document-analysis and inspection tooling runs entirely on hardware you control. No client data leaves your building.

    How local AI works →

Limited intake

Get the Design Sprint one-pager

What happens in each of the four to six weeks, what you receive, and what we need from you at the review.

This reaches us as an email and we reply to it — we are not adding you to a mailing list, because we do not run one yet. See the privacy policy.

Questions we get

How is AI actually used, and what does it never decide?

It widens and it drafts. AI generates a far larger concept set than a person would sketch in the time available, drafts requirement and test-plan structure, and cross-checks documents against each other. What it never does is decide: which concept survives, whether a risk control is adequate, whether a material is appropriate, and whether the design is ready to move on are engineering judgements made by a named person who signs for them. If a deliverable cannot be traced to a human who reviewed it, it does not leave.

Who owns the design and the IP?

You do. Anything created for you in the sprint — concepts, CAD, drawings, the prototype — is yours, and that is written into the statement of work rather than implied by a terms page. We keep only the limited licence needed to do the work you engaged us for. If you are working under an institutional or investor agreement that touches IP, tell us at the review so it is handled before anything is drawn.

What does "prototype" mean materially?

A machined part or assembly in a representative material, with real geometry and real tolerances, suitable for bench characterisation and for putting in a clinician’s hand. It is not sterile, not validated, and not for use in or on a patient. If your concept needs a material or process we cannot machine in-house, we say so at the review and quote the supplier route honestly rather than absorbing it silently.

How does this connect to the documentation and verification lines?

Deliberately. The sprint produces requirements and a test plan in the structure the V&V line scans and the documentation line builds from, so nothing is retyped or reconstructed at the handoff. You are not obliged to use either — the outputs are yours and they are standard artefacts — but if you do, the second engagement starts from a file that is already in the right shape.

Our program is sensitive. Can this run without our data leaving our building?

Yes, for the document and analysis side. Our document-analysis tooling can run on hardware you control, so drawings, requirements, and design files never leave your network. CAD and machining obviously involve moving files to us under the confidentiality terms in the SOW; what the local option removes is the AI-processing layer as a separate place your data would otherwise sit. We will tell you exactly which steps run where before you sign.

What if the concept turns out not to be worth building?

Then the sprint says so, with the reasoning written down, and that is a result worth having for a fixed fee in six weeks rather than a year. Sometimes the honest output is a different concept than the one you arrived with; sometimes it is that the requirement everyone assumed was fixed is the thing that needs to move.