Key Takeaways

• MetriCal reduced camera calibration time to 13% of the original process by enabling Kodiak to calibrate multiple cameras simultaneously

• Integration of MetriCal into production workflows allowed for automated, one-button calibration processes at contract manufacturing facilities

• MetriCal's validation capabilities helped identify calibration issues early in the process, improving the automated quality assurance at the contract manufacturer

• Responsive technical support enabled rapid problem-solving as Kodiak scaled from development to production

The Mission

Autonomous trucking represents one of the most challenging applications of self-driving technology, demanding exceptional sensor precision for safe operation at highway speeds. Class 8 autonomous vehicles rely on multi-modal sensor systems, where even small calibration errors can compromise safety and performance. For companies transitioning from development to commercial deployment, establishing production-grade calibration workflows becomes critical to scaling operations.

Kodiak AI is building the future of autonomous trucking with the Kodiak Driver, an AI-powered autonomous system designed for commercial long-haul operations and industrial applications. Founded in 2018 and based in Mountain View, California, Kodiak became the first company to deploy customer-owned-and-operated driverless trucks in commercial service in 2024. With over 3 million autonomous miles logged and daily freight delivery for partners including Atlas, J.B. Hunt, and Werner Enterprises, as well as partnerships with the Department of Defense, Kodiak operates at the forefront of autonomous trucking commercialization.

Court Hinricher, who works with the VP of Hardware on special hardware and operations projects, identified Tangram Vision as a primary calibration software partner. As Kodiak expanded from development to production, the calibration challenge became increasingly acute: how to scale up manufacturing and, at the same time, maintain sub-pixel reprojection error.

Building for Scale

Kodiak's sensor architecture includes lidar, radar, and cameras with various lenses integrated into their proprietary SensorPods. Kodiak SensorPods are designed for both quick field replacement and simplified OEM integration. To scale manufacturing and maintain sensor calibration requirements, Kodiak needed a calibration procedure that could guarantee accuracy while also delivering speed of throughput.

Kodiak's build-versus-buy analysis determined that leveraging the intrinsic calibration expertise of Tangram would accelerate the development of their vehicle commissioning process while maintaining the precision Kodiak’s system demanded.

From Development to Production Line

Kodiak's calibration journey progressed rapidly from initial development testing to full production integration. The team started by validating MetriCal's performance on their camera systems with strict in-house testing. Once they were certain of the calibration quality MetriCal delivered, the Kodiak team used the calibration suite as a basis for a one-button process that their contract manufacturers could adopt.

This automation was critical for Kodiak's production strategy. Rather than requiring specialized expertise at the contract manufacturing facility, MetriCal enabled a streamlined workflow that production technicians could execute reliably. In fact, MetriCal's detailed reporting enabled Kodiak to validate calibrations from external suppliers, demonstrating the value of having robust validation tools.

"We appreciate the back-and-forth with the [Tangram Vision] team; they’ve been helpful throughout the calibration development process."

— Court Hinricher

The impact of MetriCal on Kodiak's operations has been substantial. The team achieved dramatic improvements across multiple dimensions. Kodiak reduced camera calibration time to 13% of the original duration using their latest processes. This improvement came from MetriCal's ability to process multiple cameras simultaneously, transforming what was once a sequential bottleneck into an efficient operation. In turn, these accurate calibrations delivered measurable improvements in Kodiak’s core perception technology, directly impacting the Kodiak Driver’s performance.

Technical Support When It Counts

Beyond the software capabilities, Kodiak valued the collaborative relationship with Tangram Vision's engineering team. Court emphasizes, "We appreciate the back-and-forth with the [Tangram Vision] team; they’ve been helpful throughout the calibration process."

This support proved particularly valuable as Kodiak navigated complex technical challenges, including optimizing narrow field of view camera calibrations and ensuring consistency across different production environments. Direct access to domain experts accelerated problem-solving and enabled Kodiak to maintain their aggressive production timeline.

Jeremy Steward (left) and Brandon Minor (center) of Tangram Vision, visiting with Court (right) and the Kodiak team at CES 2026

Looking Forward

As Kodiak continues scaling operations with plans to deploy many more autonomous trucks with their logistics partners, calibration infrastructure remains a foundational capability. The company is already planning for expanded calibration facilities at its contract manufacturer, supported by MetriCal’s production line calibration abilities.

Court's team is focused on continuous improvement: increasing calibration throughput, reducing failure rates, and optimizing processes for different camera modalities. The flexibility of MetriCal Enterprise, including unlimited calibrations and dedicated engineering support, positions Kodiak to scale without calibration becoming a constraint. Kodiak can now focus their engineering talent on what matters most: building the safest, most capable AI-powered autonomous driving technology in the industry.