Key Highlights:

If you're short on time, here are the main points from Mohammad Musa's insights:

• Early Career: Started in the semiconductor industry, then worked on game physics engines for titles like Half-Life, Uncharted, and Resident Evil.

• Google Experience: Spent six years at Google managing products like Gmail and Drive.

• Founding Deepen AI: Bootstrapped Deepen AI into a multimillion-dollar business focusing on sensor fusion and calibration for self-driving cars.

• Major Clients: Works with 7 of the 10 largest car manufacturers, including BMW, Mercedes, Nissan, and Bosch.

• Hybrid Approach: Uses both AI automation and human checks to ensure safety.

• AI and Jobs: Believes that increasing AI capabilities will create more human jobs due to the growing amount of data.

• Future Plans: Aims to expand into agricultural robots, medical imaging, and AR/VR applications.

The Journey to Founding Deepen AI

Mohammad's path wasn't straightforward. After engineering school, he worked with semiconductors. Then he moved into the gaming industry, working on physics engines at Havok for games like Half-Life and Uncharted. This gave him deep experience with 3D graphics and immersive tech.

He later spent six years at Google, managing product strategies for Gmail, Drive, and Calendar. In 2016, he won Facebook's Launchpad competition for innovation in AR/VR. However, he realized that the hardware for virtual reality wasn't ready for everyone yet. So, he turned his focus to self-driving cars, which faced similar challenges in spatial computing.

Core Technology and Innovation

Deepen AI specializes in two key areas:

1. Sensor Fusion & Calibration

   • Combining Sensors: Merges data from cameras, LIDAR, radar, and thermal sensors.

   • Accurate Alignment: Ensures that all sensor inputs line up correctly.

   • Real-Time Adjustments: Keeps the car safe by adjusting calibrations on the fly.

   • Factory Calibration: Provides calibration services for new cars right off the production line.

2. AI-Assisted Annotation with Human Validation

   • Hybrid Approach: Uses AI to automate tasks but includes human checks.

   • High Accuracy: Maintains the precision needed for safety-critical applications.

   • Scalable Solutions: Handles increasing amounts of data without losing quality.

   • Adaptable: Adjusts to new sensor technologies and different types of data.

Future Vision

Beyond the automotive industry, Mohammad sees opportunities in:

• Agricultural Robotics: Helping to produce sustainable food for a growing population.

• Medical Imaging: Improving how we analyze medical scans.

• AR/VR Applications: Enhancing virtual and augmented reality experiences.

• Digital Twin Technology: Creating virtual copies of real-world industrial settings.

Interview Deep Dive

- On Building Deepen AI

Why did you pivot from AR/VR to automotive?

I participated in a Facebook competition in 2016 and was one of the winners. But I quickly realized that augmented and virtual reality were still in the early stages. Even now, the technology isn't fully ready for mass adoption.

- On Technology

Can you explain spatial-visual intelligence in simple terms?

Think of 2D data from cameras—those are just pixels with colors. When you add 3D, you're looking at depth, not just flat images. By combining different sensors, like cameras and lasers, we get a fuller picture of the world around us.

How does your auto-calibration technology enhance vehicle safety?

If your car's side mirror gets bumped and the camera moves slightly, our software can adjust for that change. You won't have to get it fixed right away, and your car can still keep you safely in your lane.

- On AI and Human Integration

What's your perspective on AI replacing humans?

I don't think we'll ever fully replace humans. At Google, we had many people working on data labeling. Even as we got more automation tools, the amount of data grew, so we needed even more people.

How do you approach the balance between AI and human validation?

Sometimes, AI can't handle certain tasks, and we rely entirely on humans. Most of the time, we have both AI tools and humans working together to meet the quality our customers need.

- On Manufacturing and Scale

How do you handle cost considerations in automotive manufacturing?

Car companies are very sensitive to costs. Even a difference of one cent on a part matters when you're making millions of cars. Sometimes, you might sacrifice a bit of accuracy to make things cheaper. Engineers have to make these tough decisions.

- On Future Vision

What opportunities do you see beyond automotive?

I'm particularly interested in agriculture. With nearly 10 billion people to feed, ensuring everyone has healthy food is a major challenge. We need to figure out how to grow and harvest food efficiently without harming the planet.

How do you see sensor technology evolving?

Each new generation of sensors changes how we calibrate them. New sensors bring in information we didn't have before. Some might make calibration easier, others harder. We have to adapt our methods accordingly.

- On Entrepreneurial Journey

What's your advice for technical founders?

You have to build it right—that's engineering. You have to build the right thing—that's product management. And then you have to sell it. Many engineers underestimate the importance of good salespeople. Great sales and marketing can deliver the value you've built.

How did you get your first customers?

Our first customer was an ex-Googler. We have a network called Zooglers—ex-Google employees. Mercedes reached out to us after finding us through a conference or our website. Having a solid online presence is crucial. Invest in SEO and marketing; it really makes a difference.