Parametric CAD bridge for topology optimization workflows

The specific moment: an engineer runs topology optimization in Altair Inspire or a similar tool, gets a result that shows a promising organic structure, and then discovers they have almost no ability to parameterize it, constrain it to manufacturing rules, or export it as anything other than a low-quality STL that's nearly useless for production handoff. The complaints are direct — 'algorithm for transforming TO results to CAD surfaces is not effective', 'low quality STL output during implicit modelling', 'modeling function was difficult to parameterize and often required using CAD and importing a part rather than tightly controlling the base workpiece'. This isn't fixable inside the existing tools because topology optimization is the product; the CAD output is an afterthought bolted on to complete a marketing checklist. The simulation vendor has no incentive to build a full parametric CAD engine when their value proposition is 'run the optimization, trust the solver.' The result is that engineers who are supposed to be accelerating design cycles are actually doing more manual work than if they'd just modeled the part by hand.

The gap is structural: TO output is implicitly defined geometry, and converting that faithfully to a parametric solid with editable features and manufacturing constraints requires a completely different technical stack from what a solver company maintains. Nobody at the solver vendor has the incentive to staff that team.

This is a business because every topology optimization project — and there are more of them every year as lightweighting pressure grows in aerospace, auto, and medtech — generates this exact bottleneck. The cost is concrete: either a full CAD rework by a senior engineer (hours of labor per iteration) or a low-fidelity STL handoff that delays manufacturing sign-off. Teams doing more than a handful of TO projects per month will pay to eliminate that rework loop.

What to build

Build a file-in, file-out conversion service and lightweight editor that takes implicit or mesh-based TO output (from Inspire, OptiStruct, or similar), reconstructs it as a parametric solid with user-adjustable feature controls and wall-thickness constraints, and exports clean STEP ready for a CAD PDM system.

Where to start

Target medtech device companies first — they run TO for implants and brackets, they have strict manufacturing tolerances, and their engineers are already used to paying for specialized preprocessing software because FDA workflows require documented geometry control at every step.

The hard part

Reconstruction quality on organic TO geometry is genuinely hard — you need to decide early how much you're willing to approximate versus preserve exact simulation-derived topology, and that trade-off will alienate one customer segment no matter which direction you go.

How it makes money

Usage-based pricing per conversion job (suitable for consultancies with variable volume) plus an optional annual team subscription for in-house engineering teams doing more than 20 jobs per month.