AI Researcher

About Forgis Forgis AG is a deep-tech startup based in Schlieren (ZH), affiliated with the ETH AI Center. We are developing an Industrial World Model - the "brain of the modern factory": a platform that connects machines, controllers (PLCs), and robots across different manufacturers. Funded by a redalpine-led pre-seed. Currently 5 accepted publications at ICML 2026 (incl. spotlight). Your Role As an AI Researcher, you are responsible for our research agenda around self-supervised representation models and World Models for industrial systems. You drive fundamental, ambitious research – from physically grounded architectures to new paradigms in Representation Learning, in close cooperation with ETH, Oxford, and Imperial College. Areas of responsibility • You research self-supervised and JEPA-like Representation Models (HEPA family) for industrial time series and multimodal sensor data • You develop physically grounded inductive biases, causal grounding, long-horizon stability, conservation of information, as an alternative to purely generative or VLA-based approaches • You develop models for anomaly detection and predictive maintenance across machines and manufacturers and consistently drive them towards World Models • You transition fragmented individual models into a machine-agnostic foundation model (edge encoders → central server model) • You ensure efficient inference on edge and scale our pretraining pipelines on the industrial corpus • You publish at top conferences and build our scientific reputation Requirements • PhD in ML, CS, robotics, or physics • Proven track record in self-supervised learning, representation learning, anomaly detection, or time series modeling; interest in World Models is an advantage • Proficient use of PyTorch and large-scale training • Advantageous: fast ML inference (FPGA/hls4ml), data compression algorithms, and edge deployment • Independent, hands-on way of working in a startup environment; very good English skills (working language)