Inspiration
From kindergarten, children are asked what they want to be when they grow up. While answers tend to be simple but broad at first—police officer, firefighter, rescue worker, or pilot—children are eventually told to be realistic and specialize in one thing. Such specialization provides focus, but only at the cost of siloing expertise and preventing cooperation.
Bring this same attitude into the engineering world, and you can kill entire projects. This is the root of the problem we set out to solve: How do we enable individual expertise while maintaining all parts as elements of a single, cohesive system?
But why? Why would we spend so much time building this? Because we aren't approaching this problem from the outside. We are engineers ourselves. We founded our own engineering team and built real machines, and we have personally experienced projects going through iteration after iteration only to fail because one person had a measurement wrong and nobody caught it. We've seen firsthand how difficult it is to keep an entire team together when everyone is working on different parts of the same machine.
Inventivus came from that experience. We built it because we wanted the tools we needed to build better machines ourselves.
What it does
Inventivus provides each person or engineer within a system with a real-time, comprehensive picture of the total system.
One could think of this as every soldier, from the basic infantry fighter to the unit leader, having complete access to a full, real-time picture of the battle—or project—as it unfolds.
Built on an AI backend system called Athena, Inventivus maintains an evolving model of an entire project. Yet this is fundamentally different from simply storing project information.
Instead of treating every CAD file, code file, measurement, design decision, and conversation as separate pieces of information, Athena connects them into a model of the machine itself. It understands not only what each piece contains, but how those pieces relate to one another and how changes propagate through the system.
When an engineer changes something, Athena can reason about that change in the context of the entire project. It can identify inconsistent dimensions, conflicting components, missing information, violated requirements, or changes that no longer make sense based on earlier decisions.
If something is wrong, or Athena has a suggestion or idea, Athena can bring it to the team's attention while they are working, help determine what needs to change, and update the project's shared context.
If the team wants to try something new, Athena can help them reason through the idea, prototype it, and develop it alongside the rest of the machine.
The goal isn't to replace engineers. It's to make sure engineers have the full context they need to do their jobs, regardless of their specialization, and to stop the constant checking of scattered files and fragmented context.
An electrical engineer should be able to understand how their changes affect the mechanical system. Inventivus gives the entire team a shared understanding of what they are building, so the project doesn't slowly drift apart as it evolves.
How we built it
We built Inventivus entirely from the ground up using an Electron JS, Typescript, CSS, and HTML.
All of the programming was done completely by hand, and took roughly 2 years to complete. Debugging of the code and de-spagghetifying of the code was partially done by services such as Claude Code, but 99% of the coding was done by hand.
The backend is an agentic harness called Athena that can be run with local models, or through the Inventivus AI provider system. This allows Athena to reason about relationships between different parts of the project. When one component or decision changes, the system can consider how that change affects the rest of the machine rather than treating it as an isolated modification.
Workspace file paths are stored in the Inventivus database/context, enabling the Athena system to access them at any given time. All user data as well as project data is stored locally on the machine, in the paths that they normally would be stored at unless otherwise specified, to avoid overcrowding the Inventivus app files.
The system is designed around the workflow we experienced while building our own machines. Instead of forcing engineers to constantly search through different files, ask other team members for context, or manually check whether a change has caused problems elsewhere, Inventivus aims to maintain that context continuously.
Challenges we ran into
The biggest challenge was that engineering projects are inherently interconnected.
A single measurement can affect a component, which can affect another component, which can affect a requirement, which can ultimately affect the entire machine. Important information also exists in many different forms—from CAD files and code to measurements, design decisions, and conversations.
This meant that simply storing project information wasn't enough. We needed to think about how all of that information relates to everything else and how those relationships change as a project evolves.
Another challenge was making sure Inventivus solved a real engineering problem rather than becoming another productivity application. We wanted the system to be built around problems we had actually experienced ourselves.
We've seen firsthand how a small mistake or missing piece of context can survive through multiple iterations and eventually cause an entire prototype to fail. Designing around that experience became one of the central challenges—and motivations—behind Inventivus.
Accomplishments that we're proud of
We're proud that Inventivus came directly from our own experience as engineers.
We didn't start by trying to make another productivity app or chase quick cash. We started by building our own engineering team and real machines, and we experienced the problems of fragmented information ourselves.
We're also proud of the concept behind Athena: an AI system that doesn't simply store or retrieve information about a machine, but builds an evolving understanding of the machine itself and the relationships between its components. Athena was particularly difficult to build, due to things like low model context, and live speak, which is difficult to implement.
Most importantly, we're proud of the fact that the app has come to fruition at all. We accomplished something incredibly difficult that could have a real positive impact on a community that we love.
What we learned
We learned that engineering problems are rarely isolated.
A decision made by one person can have consequences for dozens of other parts of a system, even when those consequences aren't immediately obvious. Specialization is necessary, but specialization without shared context can cause teams to lose sight of the system as a whole.
We also learned that AI becomes significantly more useful it's not in your way, and not there to Athena isn't intended to simply answer questions about a project. Its purpose is to understand how the project's information connects, how those relationships evolve, and what those changes mean for the rest of the system.
Most importantly, we learned the value of building around problems that have a real impact. Inventivus exists because we encountered this problem ourselves, and using it with our team has shown us how much of a positive impact a couple engineers with a dream can have.
What's next for Inventivus
Our goal is to continue developing Inventivus into a system that can even better support an engineering team throughout the entire process of creating a machine.
We want Athena to become increasingly capable of understanding project context, identifying problems before they become failures, helping engineers explore new ideas, and assisting teams as they prototype and iterate.
Ultimately, we want Inventivus to make it possible for engineers to specialize without becoming isolated. We want to scale Inventivus to a real product that other engineers around the world can use to their benifit as we did.
We aren't trying to replace engineers. We're building a system that gives them the context they need to work together.
Inventivus is built by engineers, for engineers—to help teams build better, build faster, and turn ideas into real machines that change the world.
Log in or sign up for Devpost to join the conversation.