Postdoctoral Researcher (Postdoc) - Artificial General Intelligence (AGI) Beyond LLMs (ARC-AGI-3 Focus) (m/f/d)

Postdoctoral Researcher at TIB – Leibniz Information Centre for Science and Technology in the AI for Science Research Group (Prof. Dr. Sahar Vahdati). We focus on Artificial General Intelligence (AGI) beyond LLM-centric paradigms, using ARC-AGI-3 as a primary reference framework to define and evaluate progress—especially in robust generalization, abstraction, and problem solving beyond pattern matching.

Core fields: AGI & reasoning, machine learning, cognition-inspired AI (phenomenology of cognition: reasoning, agency, meaning/experience), evaluation & benchmarking, robust/OOD testing, game-based or interactive environments, reproducible ML engineering.

Key tasks & responsibilities:

Develop and implement ARC-AGI-3-aligned methods and run systematic evaluations in close collaboration with the team.
Translate research hypotheses into testable computational models, experiments, and measurable outcomes—exploring both generative AI and alternative approaches beyond standard deep learning.
Design reproducible experimental pipelines (baselines, ablations, robustness & out-of-distribution tests), including game-like environments where appropriate.
Plan and structure research work (milestones, progress reporting, iterative refinement based on empirical results and feedback).
Build and maintain reproducible research code in Python (version control, automation of training/evaluation workflows, experiment configuration), including the use of AI-assisted development tools.
Communicate results through documentation and presentations; contribute to scientific publications and open research artifacts (code, configurations, experimental setups).