What engineering intelligence really looks like: from invention to adoption

The CMOS image sensor I helped develop now sits inside billions of devices worldwide. What most people don’t realise is that one of the biggest challenges wasn’t inventing it – it was finding a path from the laboratory to widespread adoption.

I learned that lesson while working as an engineer at NASA’s Jet Propulsion Laboratory (JPL) in the 1990s. Our team was developing a new type of camera chip for space missions. In the process, we created several imaging innovations that led to a new type of solid-state imager: the CMOS Active Pixel Sensor (CMOS-APS). We quickly realised that its advantages extended far beyond space exploration. Its lower power consumption, smaller size, and lower manufacturing cost made it attractive for countless terrestrial applications.

I co-founded Photobit to commercialise the technology and bring it to market. As CEO, one of my first responsibilities was securing a licensing agreement with the California Institute of Technology, which controlled the intellectual property for our inventions. Securing an exclusive license gave Photobit the necessary foundation to attract investment and build a company around the technology. Intellectual property created the initial bridge between research and commercialisation. Without that bridge, inventors can struggle to bring even promising innovations to the market.

That experience reinforced something I would encounter repeatedly throughout the commercialisation process: invention alone does not lead to impact.

Over time, I came to define engineering intelligence more broadly. Beyond solving technical problems and optimising performance, engineers must recognise which solutions can be adopted, scaled, and sustained beyond the laboratory. Bringing a technology into the world requires understanding customers, manufacturing constraints, industry standards, supply chains, business models, and market timing. The engineering challenge is often only the first challenge.

My team at Photobit initially believed we needed to build and sell a complete camera in order to demonstrate the value of our image sensor technology and help potential customers understand its capabilities. Because camera development was not our core expertise, however, we spent considerable time, money, and engineering talent creating a standalone camera product.

The camera worked. But we never sold it commercially.

The problem was that we had confused demonstrating a technology with building a commercial product. We underestimated what it would take to source components, establish manufacturing processes, develop distribution channels, support customers, and bring a finished product to market.

Fortunately, this effort represented only a small part of the company’s work. Photobit continued to grow despite this and other early mistakes. But it reinforced an important principle: Engineering intelligence includes understanding which problems you should solve and which problems you should leave to others.

For startups operating with limited budgets, that distinction becomes critical. A new technology often creates dozens of possible product opportunities. Engineers naturally want to pursue many of them simultaneously. Successful commercialisation requires the opposite approach. Teams must identify the minimum viable product and remain disciplined enough to stay focused on it.

Many startups fail not because their technology is inadequate, but because they spend too much time and money perfecting features that customers neither need nor value.

We learned this lesson the hard way with our first commercial image sensor. Like many engineers, we assumed customers would want the highest-performance product we could create. We designed our first CMOS image sensor with the largest number of pixels we could fit within a constraint imposed by an early corporate investor. The result was a 512 x 384 pixel sensor.

We were proud of the sensor, but we could not sell it.

The reason was simple. Existing camera manufacturers had already designed products around standard CCD sensor formats. They needed a compatible replacement, not a technically superior but incompatible alternative.

See Also

The next frontier of AI isn’t about replacing people – it’s about empowering them

To enter the market, we ultimately developed a much smaller 256 x 256 pixel sensor that matched existing industry expectations. Although it offered lower resolution than our original design, customers could integrate it into existing systems more easily.

That experience fundamentally changed how I thought about innovation. My team had focused on maximising technical performance, while our customers were focused on minimising disruption. Markets often reward compatibility before they reward technical perfection.

This insight extends far beyond image sensors. Engineers naturally focus on increasing performance, expanding capability, and pushing technical boundaries. Customers, however, evaluate technologies within the context of existing systems, manufacturing processes, budgets, and workflows. As a result, they often choose solutions that fit existing products and standards over technically superior alternatives that require significant change.

The image sensors we developed for space eventually became integrated into billions of computers, cars, smartphones, and medical devices, but there was nothing inevitable about that outcome. Turning an invention into a successful product required supply chain optimisation, customer adoption, market acceptance, and mastering countless other non-engineering aspects of commercialisation.

Creating a breakthrough is an extraordinary achievement. But knowing how to bring that breakthrough into the world is just as important.