The Robot That Reads the Blueprint
Before Tessa Lau ever thought about autonomously printing construction layout plans, she was watching contractors scrawl chalk lines on a concrete floor during a Bay Area home renovation - armed with a tape measure, a chalk box, and a surprising number of errors. The imprecision wasn't careless. It was the industry standard. And it struck her as absurd.
Today, Lau runs Dusty Robotics from Mountain View, California, where her team has built the FieldPrinter: an autonomous mobile robot that takes a BIM (Building Information Modeling) file and prints construction layout - walls, columns, electrical conduit paths, all of it - directly onto the concrete substrate with 1/16-inch accuracy. That's roughly the width of a matchstick head. On a chaotic, multi-contractor jobsite. At scale.
The numbers that tell the rest of the story: $69.5 million raised, 200 employees, and over 100 million square feet of layout printed across thousands of buildings. Major contractors like DPR Construction, Turner Construction, and Performance Contracting are among the clients. Layout-as-a-Service - the model Lau pioneered - means customers don't buy a robot. They buy an outcome.
What 1/16-Inch Means in the Real World
Manual chalk lines: 1/4" to 1/2" error. FieldPrinter: 1/16".
Eleven Years at IBM, Then a Hard Pivot
Lau graduated from Cornell University with degrees in computer science and applied engineering physics before earning her PhD in machine learning from the University of Washington in 2001. Her dissertation focus - human-computer interaction layered on top of machine learning - would define the throughline of everything that followed.
For 11 years at IBM Research, she worked on business process automation and knowledge capture: teaching computers to understand the implicit steps inside complex workflows. It was important, intellectually demanding work. It was also, by her own admission, not quite the wide-open field she was looking for.
The break came through Willow Garage, the now-legendary Silicon Valley robotics research lab that spawned a generation of robotics companies and researchers. Lau joined as a research scientist and immediately felt the difference between building software abstractions and building things that moved through the physical world. She was hooked.
"I wanted to get back to a wide-open field with numerous possibilities."
Tessa Lau on leaving IBMFrom Willow Garage, she co-founded Savioke in 2013, where she served as CTO. Savioke's product - the Relay robot - delivered items through hotel hallways and high-rise corridors. She and her team deployed over 75 robots into the hospitality industry, navigating real-world service environments with human guests, narrow corridors, and elevator banks. When she left in 2018 to start Dusty, she had spent five years putting robots into places humans had always managed alone.
The House That Built the Company
Lau has been candid about what sparked Dusty Robotics: a home renovation in the Bay Area. She watched contractors doing manual layout with tape measures and chalk, noticed the inevitable errors, and started asking why nothing better existed. She then began visiting active construction sites to understand the industry's workflow - not as a tech tourist but as a systematic problem-diagnosing researcher with a PhD in machine learning.
What she found was a sector worth trillions of dollars globally that had been insulated from technological disruption by complexity, fragmentation, and the sheer physical nature of the work. Layout - the process of marking where everything goes before building begins - sat at the beginning of almost every construction sequence and propagated any errors downstream. A wrong line at layout became a misaligned wall, a code violation, a rework invoice.
"People come to us looking for a robotic solution to increase productivity, and I think they're thinking of the wrong thing. They should be focused on quality."
Tessa Lau, ICBA Construction Innovation Summit, 2024She co-founded Dusty Robotics in April 2018 with Philipp Herget, who had been the robotics hardware lead at Savioke. The founding team brought together machine learning, robotics hardware, and a deep respect for the realities of construction - a combination that turned out to be rare and valuable.
In the first year (2019), Lau made a deliberately anti-stealth move: Dusty offered free print jobs to construction companies and gathered monthly feedback. She wasn't hiding the product until it was perfect. She was learning, month by month, what perfection actually meant to a general contractor.
The First $5,000 and a Google Search
In early 2020, Dusty completed its first paid project - a medical office building in the Bay Area. When the check was due, Lau realized she had never written an invoice before. She Googled it. The company that had invented a new category of construction automation, led by a researcher who had published on AI for over two decades, had to look up the most basic transaction document in business. The company now has 200 employees.
17 Conflicts Per Contractor, Per Day
Lau has a habit of leading with data, and one figure she has cited repeatedly speaks to why the FieldPrinter does more than just save time on layout: on a typical jobsite, the average contractor has 17 interactions per day with people from other companies, and roughly half of those interactions involve conflict.
"That conflict happens because they don't have the information they need or are working on different versions of information and come to loggerheads with a different view of reality. With a single view of reality printed under your feet, it leads to much less conflict and a safer jobsite."
Tessa Lau, October 2024This is the deeper argument behind Dusty's product: printed layout isn't just a faster way to do what chalk lines did. It's a communication artifact. A single version of the design, physically instantiated on the floor, becomes a shared reference that reduces the information asymmetries that generate conflict, rework, and schedule delays. Multi-trade coordination - the hardest problem in construction scheduling - gets marginally less hard every time every contractor is literally standing on the same drawing.
Selected Quotes
"Don't be in stealth. Stealth is stupid."
On advice for robotics founders"You gotta be able to say the name in a loud bar."
On naming Dusty Robotics"Our customers don't buy a robot. They need an outcome."
On the Layout-as-a-Service model"Robots are the great equalizer on a job site. You don't need to have great physical strength."
Blueprint 2023"We needed 1/16 of an inch accuracy - so we invented new localization technology using construction site total stations."
On the technical challenge"I want my robots to be functioning members of society."
On her broader philosophyCareer Arc
Academic Background
Cornell University
BA & BS - Computer Science and Applied Engineering Physics
University of Washington
MS & PhD - Computer Science (Machine Learning with HCI emphasis)
Awards & Achievements
- Fast Company Most Creative People in Business (2015)
- Silicon Valley Business Journal Woman of Influence (2017)
- Inc. Magazine Top 5 Innovative Women to Watch in Robotics (2018)
- OZY Magazine Rising Star designation
- Board member, CRA-W (Committee for Women in Computing Research)
- Advisor at Cantos Ventures; Venture Partner at NextGen Venture Partners
- Keynote speaker, IEEE ICRA 2025
- Raised $69.5M including $45M Series B for Dusty Robotics
- FieldPrinter technology used across thousands of commercial buildings
Things Worth Knowing
Dual-degree Cornell grad - computer science AND applied engineering physics. The hardware intuition wasn't an accident.
Her home renovation - watching manual chalk layout errors on a Bay Area project - was the direct spark for a company now working on thousands of commercial buildings.
She went from hotel hallways (Savioke delivery bots) to skyscraper floors (Dusty FieldPrinter) without changing industries - both are fundamentally about robots navigating human-operated environments.
To hit 1/16-inch accuracy, Dusty invented new localization technology using construction site total stations - standard survey equipment repurposed as robot positioning beacons.
Layout-as-a-Service (LaaS) was Lau's deliberate choice: customers never buy the robot. They subscribe to outcomes. The hardware stays Dusty's problem.
Eleven years at IBM Research gave her something most robotics CEOs lack: a systems-level intuition for enterprise workflow automation at scale.
Tessa Lau On Camera
After the Printer, Mentorship
Lau has been explicit about what she wants to do after Dusty: mentor the next wave of robotics founders. She occupies a rare perch - someone who has built robots for hospitality, raised a Series B for construction automation, and held research positions at IBM, Willow Garage, and the University of Washington. That experiential stack is not common. She is already serving as advisor at Cantos Ventures and Venture Partner at NextGen Venture Partners.
The broader ambition is structural: Lau has articulated a vision for construction becoming a precision, data-driven industry - one capable of delivering the buildings the world needs without the waste, rework, and schedule overruns that currently define it. The FieldPrinter is one piece of that. The floor-level data it generates, and the conflict-reduction it enables, points toward a construction site that operates more like a manufacturing line and less like an improvisation.
She is not interested in incremental. Her career is a sequence of hard pivots into harder problems. What comes after 100 million square feet is presumably not a slower robot.