What Problem Are We Solving?

‍

Manufacturing defects on surface-finished parts are a $10 billion annual problem in the United States alone. High-volume production methods like injection molding and casting often induce surface-level defects due to mechanical issues. Manufacturers aim to detect these defects as quickly as possible—both to prevent defective parts from entering assemblies (and requiring costly rework) and to identify root causes tied to mechanical production issues.

• 1–15% of injection-molded parts contain surface defects, resulting in wasted materials, lost revenue, and missed production targets.
• Current defect detection systems rely on rigid, hardware-dependent solutions that are slow, costly, and difficult to adapt to new parts.
• Products requiring surface finishing are often produced in high-cost-of-living areas, where manual inspection is expensive, tedious, and prone to human error.

Beyond cost, defects in surface finishing have far-reaching consequences:

• Environmental Impact: Surface finishing extends part durability, reducing waste and supporting sustainability goals. Defects directly undermine these benefits by creating more scrap and shorter-lived products.
• Safety and Reliability: Industries like automotive, aerospace, and medical equipment depend on defect-free parts to ensure safety and performance. Even minor flaws can have catastrophic consequences.
• Economic Competitiveness: With surface finishing contributing $10.7 billion annually in revenue and supporting over 167,000 jobs, defects slow production, weaken competitiveness, and reduce profitability in this vital sector.

Manual inspection not only fails to meet modern manufacturing demands but is also an incredibly boring job for workers.

Bucket Robotics was built to solve these challenges with software-first innovation and a vision for defect-free manufacturing. By deploying AI-powered, flexible systems, we're empowering manufacturers to detect defects faster, reduce waste, improve quality, and drive efficiency—paving the way for a sustainable and competitive manufacturing future.

Our Founding Team

• Led by engineers from Michelin, Uber ATG, Argo AI, and Stack AV, with deep experience in robotics and Level 4 on-road autonomy.
• Matt Puchalski (CEO): Started as a reliability engineer and worked his way up to reporting directly to the CTO, eventually becoming the integration lead responsible for building and deploying autonomous vehicles on public roads.

‍

Key Achievements

• Scaled Stack AV from zero to five autonomous trucks in just five months.
• Managed reliability for over 100 robots across the US and Germany at Argo AI, ensuring performance at scale.
• Holds four issued patents and several pending from work in self-driving technology. (See patents here: Link)

‍

From Idea to Impact

• Accepted into Y Combinator's Summer 2024 batch with just an idea, highlighting the promise of our vision and the strength of our approach. We applied after the official deadline with the desire to change how robotics and automation approach vision systems.
• Moved to San Francisco from Pittsburgh with 10 days' notice and built our first robots and camera systems from a living room in Mission Bay.
• Raised funding from incredible investors with deep expertise in AI, robotics, and manufacturing, fueling our mission to revolutionize defect detection. This includes Soma Capital, Pioneer Fund, Red/Blue Ventures, 468 Cap, Nomad Cap, and incredible angels.

How We Solve It

Bucket Robotics combines cutting-edge AI with a deep understanding of manufacturing challenges to deliver faster, more flexible, and more accurate defect detection systems. Our approach is built on three core pillars:

1. Software-First Innovation

At Bucket Robotics, we believe the key to transforming defect detection isn't new hardware—it's smarter software.

• Using synthetic data derived from CAD, our vision models are trained to detect defects in hours rather than weeks, eliminating the need for expensive, time-consuming real-world image collection.
• Our software dynamically adapts to new parts, machines, and production lines, enabling manufacturers to scale without the rigidity of traditional hardware-dependent solutions.

2. Seamless Workflow Integration

We know manufacturing doesn't operate in silos. That's why our defect detection systems are designed to integrate seamlessly into existing workflows:

• Our systems seamlessly integrate with existing cameras and conveyor belt systems, or we can provide turnkey solutions—from AI-powered cameras that tie directly into automation workflows to fully robotic systems that handle defect sorting and rejection.
• Tailored detection models ensure compatibility with individual facilities, machines, and products, reducing downtime and improving operational efficiency.

3. Versatile Applications Across Industries

Bucket Robotics is solving problems at the heart of the $10 billion annual defect issue, starting with plastics and expanding into other critical sectors:

• Injection Molding and Casting: Addressing surface-level defects that impact 1–15% of parts in these processes.
• Machined Metals and Painted Parts: Partnering with manufacturers to tackle challenges in finishing, coating, and precision production.
• Aerospace and Beyond: We've won an SBIR grant for non-destructive surface inspection on aircraft, demonstrating the flexibility and reliability of our system in high-stakes applications.
• ‍

Real-World Impact

By revolutionizing surface defect detection:

• We reduce waste, improve product quality, and save manufacturers millions of dollars annually.
• Our system empowers manufacturers to respond to defects 50x faster than competitors, ensuring they stay ahead in a global economy where speed and quality are paramount.
• From the factory floor to the finished product, Bucket Robotics is helping manufacturers deliver safer, longer-lasting, and more reliable goods to the world.

So, Why "Bucket?"

Matt's an avid gardener (to the point he wrote a book about it), and early on, he wanted to build an autonomous 5-gallon bucket that follows you in the garden. When applying to Y Combinator, a name was needed—so after searching through available domains, "bucket.bot" was the best option at the time. The assumption was that YC would help figure out a better name later, but it stuck.

The gardening bucket robot might still happen someday.

‍

‍