Building the engine of American infrastructure.
I lead robotics, software, and manufacturing at Reframe Systems, where we build software-defined factories outside Boston.
We need to learn to love building again in the US. We are short on homes, power plants, data centers, and the factories that make all of it. Since 1970, US manufacturing productivity has grown about ninefold while construction productivity has fallen, roughly 1% a year. And about 41% of the construction workforce is expected to retire by 2031. Housing and public works will keep getting more expensive unless we change how we build.
We build software-defined factories, and the structures that come out of them.
I co-founded Reframe in 2022 with Vikas Enti (CEO) and Aaron Small (Head of Operations and Digital AI). I lead Manufacturing, Software, and Robotics.
Every part we build starts as a released design model. Our software turns it into machine files, robot programs and work instructions, and schedules it on the line. Nobody redraws anything by hand. Our robotic cells frame wall panels every day in production.
FAB1 is our new factory in Billerica, MA. We got the keys, and 70 days later the first part came off the line (we had planned for 120). Total equipment cost was under $5M. And we designed it to be copied... the next one gets the same layout and the same software, close to wherever the next customer is.
Housing is the first product. 10 homes delivered so far, from ADUs to triple-deckers, and 114 more units in the pipeline.
But the robot framing a wall does not know if it is going into an apartment or into workforce housing next to a new power plant. The same model-to-factory pipeline works for the buildings behind energy and data centers, and for the carts and fixtures the factory itself runs on. That is where this goes next.
Amazon Robotics
Technical Lead for the Sparrow Program, Amazon Robotics' flagship item manipulation system. I built vision-based robotic solutions for singulation, dense packing, item identification, and damage detection, deployed across hundreds of thousands of robotic units in Amazon's fulfillment network.
Dynamic Legged Systems Lab, Italian Institute of Technology
Built the software architecture for the HyQ and HyQ2Max quadruped robots, integrating real-time controllers, state estimation, and simulation on Xenomai Linux.
WPI-CMU DARPA Robotics Challenge Team
Manipulation Lead for the WPI-CMU team in the DARPA Robotics Challenge. I led full-body motion planning for the 30-DoF Boston Dynamics Atlas humanoid robot, coordinating a team of graduate and undergraduate students through the DRC Trials and Finals.
Worcester Polytechnic Institute, 2014
GPA: 4.0Worcester Polytechnic Institute, 2011
GPA: 3.8All patents issued while at Amazon Robotics, covering robotic grasping, end-of-arm tooling, manipulation systems, and safety. View all on Google Patents
Chaitanya Mitash, Fan Wang, Shiyang Lu, Vikedo Terhuja, Tyler Garaas, Felipe Polido, Manikantan Nambi, ICRA 2023.
Nandan Banerjee, Xianchao Long, Ruixiang Du, Felipe Polido, Siyuan Feng, Christopher Atkeson, Michael Gennert, Taskin Padir, IEEE Humanoids 2015.
Christopher G. Atkeson, Benzun P. W. Babu, Nandan Banerjee, et al. including Felipe Polido, Journal of Field Robotics.
DeDonato, M., Dimitrov, V., Du, R., et al. including Polido, F., Journal of Field Robotics 2015.
Christopher G. Atkeson et al. including Felipe Polido, 15th IEEE-RAS International Conference on Humanoid Robots (2015).
Christopher G. Atkeson et al. including Felipe Polido, DRC Finals Special Issue of the Journal of Field Robotics (2015).
