Senior Research Scientist

(m/f/d)

NETWORK ARCHITECTURE DESIGN • THEORETICAL ANALYSIS • DYNAMICS & LEARNING ALGORITHMS
You will design, train and analyze novel recurrent network architectures and learning rules inspired by biological principles.
Frankfurt am Main, Germany (On-site)

The Role

As a Senior Research Scientist at Natural Intelligence GmbH, you will not just build upon existing feed-forward ML models, but you will design and test novel recurrent architectures based on biological principles. You will explore how recurrence, oscillatory dynamics, and self-organization can drive computation. Working closely with our Infrastructure Engineers, you will translate high-level mathematical theory into scalable, functional code.

Key Responsibilities

  • Novel Architecture Design: Develop and analyze computational models inspired by the HORN model and oscillatory recurrent network architectures. You will investigate how biology-inspired architectural priors can outperform current digital AI systems.
  • Dynamics & Learning Algorithms: Research and implement biologically plausible learning algorithms, synaptic plasticity, and self-organization rules. Expect to move beyond standard backpropagation to explore alternative, biology inspired readout mechanisms and learning rules.
  • Theoretical Analysis: Analyze the state spaces of dynamical systems, focusing on metastability and wave-based information processing to understand how these networks compute.
  • Simulation & Implementation: Implement biology-inspired oscillator networks and spiking neuron models using high-performance frameworks (PyTorch/JAX). You will validate these models on our internal compute cluster.
  • Cross-Disciplinary Collaboration: Collaborate with external partners on implementing these concepts on analog hardware platforms, and work in tandem with our Infrastructure Engineers to optimize research workflows.

‍

‍

Your Profile

We are seeking senior post-doc level researchers with 4+ years of postdoctoral research experience who bridge the gap between theoretical neuroscience and modern machine learning. In short, you are a scientist who codes, driven to understand the "why" behind intelligence.

Essential Technical Requirements

  • Education: Ph.D. in Neuroscience, Machine Learning, Physics, Mathematics, Computer Science, or a related field with 4+ years of experience as postdoctoral researcher is required.
  • Track Record: A proven history of research excellence, demonstrated by peer-reviewed publications in top-tier conferences (Cosyne, NeurIPS, ICML, ICLR) or journals (Nature, Science, PNAS, etc.).
  • Mathematical Foundation: Solid understanding of the mathematics underlying machine learning (linear algebra, statistics, calculus).
  • Scientific Expertise: Deep understanding of dynamical systems, recurrent neural networks, and machine learning in general. Familiarity with oscillatory dynamics or the HORN model is a strong plus.
  • Coding Proficiency: Multi-year experience with Python and deep learning frameworks (PyTorch/JAX).

Soft Skills

  • Research Autonomy: Strong self-management skills with the ability to take a research project from conceptualization to publication and implementation.
  • Communication: Excellent written and verbal communication skills in English, with the ability to explain complex theoretical concepts to engineering teams. Knowledge of German is a plus, but not required.
  • Collaborative Mindset: You enjoy working in an interdisciplinary environment where physics, biology, and computer science converge.

‍

‍

‍