Software Engineer, Computer Vision

About the role

Monumental is automating on-site construction with robotics and software. Our mission is to make construction software-defined and autonomous - a future where beautiful, bespoke buildings go up in a single day, with minimal labour.

We're a growing team of engineers, operators and technicians from companies like Palantir, Qualcomm, Dyson, Shopify, Tesla, Meta and Google DeepMind. It's still the early days of building the company, and this is a chance to make an outsized impact on a problem that matters. From day one, you’ll be working with an experienced founding team, backed by a $32M Series B led by Khosla Ventures and Plural.

Monumental is based in the city centre of Amsterdam, in a beautiful facility with an fully equipped workshop and robot testing floor. We build the whole stack in-house - hardware and software, we're explicitly not a hybrid or remote-first company - you can't build robots or walls from home, so we're onsite five days a week.

Read more on why we believe you should spend your time working on construction, and learn about our vision https://www.monumental.co/our-vision/ and company https://www.monumental.co/about/. You might enjoy our YouTube Channel https://www.youtube.com/@buildmonumental, Based in Europe’s film on how we’re fixing Europe’s housing shortage https://www.youtube.com/watch?v=2-Q2F0Fcx4A, or this video by Andreas Klinger of what our office looks like https://www.youtube.com/watch?v=DOo8n19XEns.

About the role

We’re looking for engineers with deep knowledge of software and modern 3D computer vision. We’re building an operating system to make on-site construction possible with robotics. Our software stack allows us to do everything from 3D reconstruction of the construction site (through photogrammetry), design of custom wall structures, training and deploying vision algorithms, supply chain, path planning, and robot controls and kinematics.

Some of the challenges we're solving with computer vision:

- We make 3D reconstructions of construction sites that let us match the architect's (ideal) drawings to the real world. These reconstructions are made in challenging conditions - low or harsh light, limited space - by operators who aren't computer vision experts, and we need a fast turnaround so our robots can get building immediately.

- Our robots place bricks and mortar with sub-millimetre accuracy in a global reference frame. These are higher accuracy requirements than autonomous driving, which means we ask more of our sensors - an off-the-shelf RealSense camera doesn't cut it.

- Our robots aren't always on stable ground. Sometimes we build a canal wall from a boat, or work off a wobbly scaffold. That means compensating live, so we can still build a beautiful building.

You can read about how we control robots with TypeScript in this post by our CTO, Sebastiaan. https://buildmonumental.substack.com/p/plans-as-data If you're new to robotics, you might enjoy reading Bouke's blog post on his experience joining Monumental as a software engineer https://bou.ke/blog/robotics/.