Constraint IQ was inspired by our team's firsthand experiences in manufacturing classes where we actually design, machine, and analyze parts. We saw how small mistakes in engineering drawings, like missing dimensions or unclear tolerances, can lead to wasted material, rework, and real financial loss. When you scale that across production, it adds up fast. In fact, the American Society for Quality (ASQ) estimates that poor quality, including design and documentation errors, can account for 15–20% of annual manufacturing costs in many industries. That made us think about how we could prevent those mistakes earlier, which led to the idea of building a system that checks whether a drawing is fully constrained and truly ready for manufacturing before it ever reaches a supplier.
A major challenge we faced was integrating the frontend and backend since we split into two groups working in parallel. Getting everything to connect smoothly, especially making sure the data from the analysis showed up correctly and intuitively in the UI, was more difficult than we expected. To solve this, we focused on creating a clear pipeline between upload, parsing, analysis, and visualization, and debugged at each stage so we could trace exactly where things were breaking. We iterated by testing with real drawings, identifying weak points in parsing and analysis, and refining both the data flow and UI presentation until the system was consistent. This pushed us to think more carefully about how different parts of the system communicate and made us better at coordinating as a team. Additionally, one of the biggest things we learned was the importance of moving from quick experimentation to more intentional design, especially when building something that needs to be consistent and reliable.
Overall, the impact of this project is centered on reducing wasted time, material, and cost in manufacturing by catching issues before production ever begins. Even small improvements in drawing quality can prevent expensive mistakes, and we see Constraint IQ as a step toward more reliable, standardized engineering workflows. Looking ahead, we want to expand beyond individual parts into full assemblies, including checking how components interact, ensuring constraints between parts are fully defined, and validating whether entire systems are manufacturable as a whole.
Log in or sign up for Devpost to join the conversation.