Company description Forgis is building the intelligent layer for manufacturing plants, an orchestration platform that integrates machines across vendors, hardware types, and applications. On top of this, we bring real-time intelligence: digital engineers that learn from data, make decisions, and continuously optimize production, from configuring and validating lines to predicting failures and guiding operators. From a single interface, Forgis turns disconnected automation into a unified, adaptive system: the factory’s brain. Founding AI Researcher We're hiring a Founding AI Researcher to join our research team. The work is frontier ML research: world models, LLMs for industrial scenarios, and multimodal models across vision, sensors, and language. The most important foundation models of the next decade will not be trained on the internet. They will be trained on the data of how things are made. That data is the largest untouched substrate left in machine learning: multimodal, time‑resolved, grounded in real physics, and it lives on machines that don't communicate to each other, in plants running 20‑year‑old software, in the heads of operators about to retire. Learning from it is the next foundation model problem. If you've spent your career at the frontier and suspect it's pointed at the wrong problem, come work on one that matters. Why this role Publish freely. We publish at NeurIPS, ICML, ICLR, CVPR, CoRL. Compute. 8× H100s, no quota. More for larger experiments. Independence. Frequent sparring, flat hierarchy. Industrial data. Multimodal data from real production machines, telemetry, force, vision, documents, language (in house and partnerships) Requirements Publications at top ML or robotics conferences (NeurIPS, ICML, ICLR, CVPR, RSS, CoRL, ACL, or similar). PhD (or close to defending), or MSc with significant research experience. Strong Python and PyTorch. C++ a plus. Bonus: experience with VLMs, LLM fine‑tuning, multimodal agents, or world models. #J-18808-Ljbffr
Founding AI Researcher
