Research Scientist (roles across multiple teams) at Jobisite / employment-usa.net

This job ad has been posted over 40 days ago...

applicants

Research Scientist (roles across multiple teams)

at Jobisite in United States Of America

About Us
We are a next-gen cybernetics start-up backed by top-tier investors (Our seed round led by NEA).

Our work focuses on the intersection of robotics, machine learning, and simulation. We aim to push the boundaries of what intelligent systems are capable of. We’re interested both in cases of autonomous operation, and in cases of human-collaborative operation. Our organization operates at the frontier of human capabilities, working actively to advance them and meaningfully improve the state of robotics and human society at large.

About the Role
By applying to this role, you will be considered for Research Scientist Positions across all roles at Neo Cybernetica, you will be responsible for doing the impossible.

You will be working at the bleeding edge of embodied intelligence with a world-class team spanning artificial intelligence, robotics, and simulation. You will be surrounded by experts from diverse backgrounds and work side-by-side to take The Next Step in Artificial Intelligence.

We believe that modern deep learning systems are amazing tools. Problems of perception and language modeling that were impossible just years ago are now trivial. Foundation models, novel supervision strategies, and multimodal techniques have fundamentally redefined the space of what is possible.

Our mission is to assemble and extend these technologies into the real world. As an organization, we are engineers first and foremost, but realize that often the most interesting and challenging research arises when theory meets reality. Our research scientists are the tip of the spear - breaking new ground through which the rest of the organization can follow.

Below are the four current areas that we are focused on:

Grounded Generation (3D reconstruction). From text to image to audio data, generative AI has advanced dramatically in the past several years. We believe that the next step in this journey is grounded generation. Specifically the ability to condition generative models on top of structured, explicit data as well as complex objectives such as multi-view spatio-temporal coherence and part deconstruction. We are specifically interested in this problem as it applies to 3D reconstruction.

Imitation Learning. In a field that is characterized by so many different supervision schemes that it’s almost impossible to list them all, imitation learning stands out as a unique field. We believe that there is particular value in learning to replicate unmapped action spaces as chains of differentiable operators.

Physics-Informed Priors. Most modern foundation models are made with generality in mind. Made to incorporate minimal priors such that the same model can be applied to Atari and yodeling and angst teen journaling. We have a different approach: specifically building our models to function in the real world; where energy is conserved, space-time is coherent, and walls can’t be hacked.

Communication. If a breakthrough is never cited, was it really a breakthrough? We believe that one of the most important, and under-emphasized aspects of research is its communication with a broader audience. This includes ad-hoc evaluation and analysis on new research and directed requests to communicate complex topics to technical audiences.

We believe that the culture of a research organization is important. There are many different philosophies and approaches to making progress here. Ours is to focus maniacally on multi-disciplinary work and redefining problem statements to take advantage of the flexibility afforded to private industry. These are the core attributes that we look for across all of these roles to ensure success:

Humility. Our goal, as an organization, is to surround you with brilliant people. Nobody will be the smartest person in the room and we endeavor to build a research team that will thrive in that environment.
Pragmatism. The path to hell is paved with good intentions. There are simply too many fascinating problems in our space. We believe the core to navigating these incredible opportunities and research directions is with relentless pragmatism. What is the cheapest test we can run? What is the simplest approach to a problem that might work? What are the problems that we can push to tomorrow?
Open-mindedness. We are building this impossible. That means, by definition, that nobody knows how to do the things that we are attempting. We are all wrong, and in a handful of months we will all look like fools. Only by embracing that can we make progress.

If the above resonates with you, we would love to hear from you. At Neo Cybernetica we plan on accomplishing some truly incredible things and are looking for people just crazy enough to come along for the journey. I can earnestly say to everybody reading this: I hope you are one of those people.

I hope you’re one of the crazy ones.
Reference : Research Scientist (roles across multiple teams) jobs

» Apply now