24-04-2024
MEETING WITH CLAIRE DELAUNAY, CTO FARM-NG, EX VP NVIDIA & CO-FOUNDER AT OTTO
by French Tech SF
Introducing Claire Delaunay, a pioneer in the robotics industry. With a rich background that includes time at Google, leading the self-driving truck startup Otto, and contributing to robotics at Nvidia, Claire brings a wealth of experience in engineering and innovation. Her journey from these tech giants to reshaping farming practices highlights her passion for harnessing technology to foster positive change. At Farm-ng, she leads groundbreaking efforts in revolutionizing agriculture through robotics and AI, laying the groundwork for a more efficient and sustainable future. Claire's commitment to empowering farmers and students through innovative tools underscores her vision for a more technologically integrated world.
“I believe that success in one's career comes with a responsibility to give back to society”
FTSF: Could you share your journey in the tech industry, especially being a woman in this field?
Claire Delaunay: My journey in the tech industry as a woman has been quite a ride. I know that if it had not been for my friends at the time and the people who surrounded me, I would probably have given up a long time ago. In France, the path for women with a technical brain
like mine is not easy. There's this sense of being doomed to not succeed early on, but I refused to let that define me.
When I was in engineering school, I was one of the three women among 300 students, and I was diving into computer science, which was exciting. However, I soon realized that traditional information systems didn't satisfy my thirst for something more tangible, more intense. That's when I decided to explore beyond the classroom.
I embarked on my first internship in a hospital; however, I craved something intense, something that felt impactful. So, I took a leap and applied to work at a research station in Antarctica and a volcano on Réunion Island. I got on the volcano project, they just said, “Come on in, we need help to do our Intranet so that we can monitor the channel, come very far away”.
Once I got there, I set up my Intranet in one month. And I came to them saying: "I'm done. What do I do now?” And they said: "Well, you can come with this seismic station assembled around the volcano to start collecting data and see how it works”. And then I thought that it was exactly what I wanted to do. I love building sensors and devices that observe the real world, from which we derive data and around which we build algorithms to perform calculations. It is very important to know where the eruption is going to come from because you can predict where the lava is going to go and you can evacuate in time. Since then, I've never given up my passion for observing and calculating everything that happens around it.
FTSF: What inspired your transition into robotics following this experience, and could you share how you ended up at Google and in the US?
CD: My transition to robotics was quite a funny story. After my internship, I was told there was no full-time job for me. So right after that, I joined a weather forecasting company, because, again, that connection with weather, sensors, and understanding how things work on a global scale, all that blew my mind. For two years, I did weather forecasting, but pretty soon, they put me on the money-making machine, which was the intranet, the data warehouse, and all these payment systems for weather forecasts and everything else. At that time, I decided to build a robot myself, in my apartment.
After building my robot, I went to a robotics conference in the south of France. At this conference, I met a company called Intempora which was working on software for self-driving cars. They saw the robot and said to me: "If you want to do robotics, why don't you come and interview with us?” and that's how I got my start in robotics. It was my first job at Lendem in robotics, where we worked on a self-driving golf cart called Cyber Cab. We explored the possibilities of autonomous cars early on but realized that the economics weren't quite there yet.
After some time, I met someone special and followed him to Stanford, where he did an MBA. I arrived at Stanford and discovered that it was a Disneyland for robotic systems engineers. I wondered how it was possible that in this country, even if you're a woman, if you love technology and you're good at it, people listen to you and you can do something. For me, it was a revelation and that's how I came and stayed in the United States.
Simultaneously, I was involved in a robotics project in France called CoreBot, a three-year contest that was sponsored by the DGA. My team won by building an autonomous robot that was able to patrol and discover the territory and prevent injuries in urban conflicts. A big defense group tried to buy our robot and I thought that this was not going to work for me.
I returned to the US, where I explored my love for technology and connectivity. I put things together and what can I do? And I was like, well, I love Skype. So I built a very small telepresence robot that you can put your phone on and connect to Skype. And I did build it originally because it was fun and because my family could drive around. My robot gained traction on Kickstarter and it turns out I had an order for 500 of them, and really fast I stopped doing it in my garage and went to Taiwan to contact manufacturers.
The success of the telepresence robot opened doors. Skype reached out to learn about my approach, and Google also expressed interest. That's how I began my journey at Google, leveraging my skills in robotics and technology to make meaningful contributions.
FTSF: How did you feel transitioning from working independently to joining a large organization like Google?
CD: I recall my initial experience at Google. Back then, I wasn't very proficient in English. During my first meeting, hearing everyone talk on speakerphone was overwhelming as I couldn't understand much of what was being said. It felt daunting like I might be let go by the end of the day! However, with time, I adapted to the accents and the environment.
What struck me most was discovering the imperfections within the company. I had this perception that everything was flawless, but reality proved otherwise. I had an idea for a project at Google, but I faced skepticism and a lack of support initially. I designed a small camera that could be controlled remotely over the internet. Despite being told it was impossible and that it wouldn't work, I persisted and questioned why something that was feasible years ago couldn't be done now. It became apparent that in large organizations, opinions on feasibility often come from individuals who may not have the necessary expertise. This is a common challenge in big companies where power dynamics can hinder efficiency.
There's a saying at Google that highlights the significant impact a small percentage of engineers can have on overall value creation: "1% of engineers are creating 99% of the value." This underscores the importance of finding the right avenues and support within such environments to drive meaningful initiatives forward.
Then I transitioned to Google Robotics, which was renowned for acquiring top-notch robotics companies. However, I soon noticed a concerning trend: the new hires were often more skilled than the existing Google employees. This realization highlighted a potential challenge within the organization. After a year of this, I realized that a big company environment wasn't for me, and I wanted to return to self-driving technology and startups.
FTSF: Can you share with us your journey from Google to starting Otto? What inspired you to venture into self-driving trucks?
CD: Despite my skepticism about self-driving cars' business models, a friend of mine, Anthony Lewandowski, proposed the idea of starting a self-driving truck startup. I thought it sounded feasible since I had experience building software for such tasks. That's how I co-founded Otto and embarked on the journey of building self-driving trucks.
Everything went well in the beginning. In six months, with a small team, we made significant progress in the autonomy functions, in particular maintaining trajectory on freeways. However, I was so focused on the technology that I neglected the wider landscape. About ten months later, we were acquired by Uber. Although this acquisition brought financial benefits, it led to a lot of tension and legal problems between my co-founder and his former connections at Google.
The experience taught me valuable lessons about Silicon Valley dynamics, especially regarding money, power, and how things can turn ugly quickly. Despite the challenges, it was a significant learning experience.
FTSF: Did you end up working for Uber, and if so, how did that experience unfold? Additionally, could you share how you transitioned to Nvidia?
What happened is that I worked for Uber for about a year. And then after a year, the lawsuit started going in full swing and my partner was fired. While I was turning down a transfer offer, at the same time Jensen Huang, NVIDIA's CEO, called me and said he'd heard I was doing great things at Uber and Otto and offered me to join the team to work on robotics. As long as there was no autonomous driving, I agreed. At the time, I thought I'd never work for a big company again. But when I saw the executives there, I was blown away. I said to myself: "These people are real, they don't talk about complicated things, they love technology, and they have their feet on the ground". That's the vision I have of the company I need to join now, even if it gets big it should still be very effective in terms of innovation. And that's really what motivated me to join NVIDIA at the time.
FTSF: What did you do for them? How did the story unfold? Were these things that you saw that first day turned out to be true, and you were really happy from an engineering perspective?
CD: I started the robotics department there, leading the ISAAC project, which included various hardware components, simulations, and other elements enabling the operation of autonomous robots. It was a fantastic project that I enjoyed, and we made significant progress. However, the reality is that robotics is a relatively small market for NVIDIA. If you look at the global market as a whole, except for vacuum robots, there are only a few million robots in use. This figure is well below that of industries such as the automotive industry.
After devoting two years to robotics and familiarizing himself with its complexities, Mr. Johnson presented me with some compelling projects in autonomous driving technology, focusing on parking and low-speed maneuvers. Despite initial hesitation, the consensus after our meeting was clear: I had to accept this opportunity. So I decided to take up the challenge and spent another two years working on low-speed maneuvers, including parking and precise maneuvers.
However, after this period, I realized once again that this was not in line with my long-term goals. I needed a change, so I took a break to explore other interests, including attending Burning Man, traveling to the Himalayas, and engaging in various activities. Finally, I felt reinvigorated and ready to re-enter the workforce, as I looked forward to new challenges and opportunities.
FTSF: You later moved on to Farm-ng. What inspired you to shift your focus to agriculture and robotics in farming?
CD: After much contemplation about my next steps, I realized how vital my connection to the Earth is. Growing my own food has always been a cherished part of my life, tucked away in my little homestead. It's a passion that's never wavered.
During my reflective moments, amid adventures like Burning Man and travels, I stumbled upon a group pioneering agriculture with robotics. Seeing these robots in action sparked something in me: a desire to apply my engineering and team management skills to a field that's often overlooked. Farming, to many, may seem mundane, with slim profit margins and immense challenges. Let's face it, most farmers are caught in a cycle where their margins are squeezed between input costs and market pressures.
Despite the hardships, farmers remain dedicated and passionate. This dedication resonated deeply with me, leading me to believe that success in one's career comes with a responsibility to give back to society. I've been fortunate to encounter supportive people and resources throughout my journey, shaping my path in profound ways.
That's where Farm-ng comes in. My main goal with Farm-ng is to work closely with students and universities, particularly in areas like Salinas, where Mexican-dominated colleges thrive. We focus on training people and ensuring they have the tools to learn and envision their future, carrying this forward ten years from now. This approach stems from my own career beginnings, where questions sparked during my school days inspired me.
I strongly believe that when you aim to change something crucial in society, starting with the younger generation is paramount. While I may be too old to make a significant difference now, the younger generation holds the power to enact real change. Allowing them to sketch and imagine possibilities is key. While this doesn't guarantee success, it significantly increases the odds of making meaningful progress.
Additionally, we emphasize the need to create tools that are easily modifiable and adaptable for new practices. Currently, our development cycles in technology can take up to five years, which is quite lengthy. To address this, we developed Amiga, small robots that function similarly to large tractors but with open-source capabilities. This open approach allows experts in various fields to contribute and build solutions much faster, reducing the timeline from 5 years to 6 months or a year. It's about democratizing technology to accelerate positive change.
While this journey is challenging and will take time, my motivation isn't driven by financial gain. I'm here because I genuinely believe in the necessity of this work and my responsibility to contribute to a better future.
FTSF: Can you tell us more about the devices you're building at Farm-ng and how they contribute to sustainable farming practices?
CD: It's a very small machine, referred to as a class zero, with about two horsepower. Despite the horsepower terminology, it's fully electric. The standout feature is its adaptable frame, capable of accommodating various payloads. Essentially, it's akin to a driverless tractor, with the option to add a seat if necessary. However, our primary aim is to create the ideal tool for field-based CNC operations.
For those unfamiliar, CNC stands for computer numerical control, commonly used for precision carving in materials like wood or metals. Our vision is to transform farms into CNC fields using the AI and perception technology available today. This means carrying around a versatile tool that can act as a seeder, weeder, harvester, or compost spreader based on programmed instructions, revolutionizing farming practices.
By EDHEC GETT Students: Thank you to Astelle Daquo, Ines Jackson-Gonzalez, Eléonore Bourreille, Eva Qu, and Sibylle Thiebaut, as part of their Master in Management program at Berkeley.
Back to news