Research scientist

Kausable GmbH is looking for a research scientist.

kausable GmbH is pioneering a reasoning-first approach to AI, developing a foundation that's fundamentally different from Large Language Models. While today's models rely heavily on pattern matching and massive pre-training, they focus on logical reasoning and adaptability.

They are creating agentic AI that:

* learns in-context from just a few examples
* opens up whole new domains and skills without retraining
* generalizes across tasks and modalities

If you're an experienced researcher, smart, creative, and inspiring, you'll fit right in!

The Company

kausable has recently raised a large funding round with investors from Black Forest Labs (a pioneer in visual generation), tech experts, and a strong network of advisors and supporters.

Tasks

kausable's research and development currently revolves around generating synthetic world data, training and validating deep-learning models, creating and facilitating capable embedders for various domains and modalities. You'll have the opportunity to:

* design and scale deep neural networks pre-trained on synthetic data for active learning and dynamic task solving
* develop new model architectures and datasets for real-world reasoning
* extend synthetic-data pipelines and evaluate across tasks and modalities
* collaborate on open-source initiatives and contribute to research publications
Requirements

The start-up is looking for research scientists with a strong background in one or more of these topics:

* active learning
* meta-learning
* modeling of complex and evolving systems
* reinforcement learning
* Bayesian learning, Bayesian optimization
Recommended qualifications:
* a PhD in ML, Physics, or equivalent – or MSc with exceptional experience
* a strong grasp of causality, meta-learning, PFNs, and active inference
* the ability to work independently and think from first principles
* hands-on experience with modern ML tooling and research workflows
* an outcome-oriented mindset
Additional nice-to-haves:
* publications in NeurIPS, ICML, ICLR, ECCV/ICCV, or similar
* experience with graph-based models or synthetic data generation
* open-source contributions
Company culture

kausable is committed to 'putting science at the core of AI'—with curiosity, daring, and humanity. They:

* are scientists at heart, with a builder’s mindset
* are open to challenge, grounded in curiosity and respect
* welcome diverse perspectives and value thoughtful and open debate
* focus on outcomes and real-world impact
* create an environment of support, inspiration, and freedom for everyone to excel
Perks and benefits:
* competitive salary + equity
* conference travel
* flexible working hours with work-from-home options
* pragmatic and supportive coworkers
Tools and infrastructure:
* Weights and Biases
* Docker
* RunPod
* AWS
* a high-end laptop of your choice
Interested?

Send your favorite way to drink coffee along with your CV or LinkedIn profile! They will get back to you soon.

If selected, you'll have one to three casual interviews to get to know each other.

Note: The job posting is active, and they are accepting applications.

#J-18808-Ljbffr


Standort

Cyber Valley GmbH, Heidelberg