KIIO is building AI for AI for process reliability in chemical manufacturing and beyond. Chemical and process manufacturing plants generate enormous amounts of sensor data, yet troubleshooting remains largely reactive. Emerging problems can propagate across interconnected equipment and control loops long before conventional alarms reveal what is happening, while identifying the true root cause can still require hours or days of manual investigation. KIIO develops interpretable diagnostic technology that learns how industrial processes behave from operational data and detects when these relationships begin to change. By modelling the plant as an interconnected dynamical system, our technology identifies emerging faults, traces how disturbances propagate, and helps engineers understand where a problem originates, what may be causing it, and how to respond, without requiring historical examples of every fault. Our goal is to give process engineers and operators the intelligence they need to intervene before deviations become downtime, equipment damage, or quality losses. Role Description We are looking for a Founding Research Engineer to help turn KIIO's research into a robust industrial product. You will work at the intersection of machine learning, dynamical systems, fault diagnosis, and industrial process data, developing algorithms that learn from multivariate sensor data and reason about complex interconnected systems. Your work will span the full research-to-product cycle: formulating problems from real industrial use cases, developing and evaluating new methods, validating them on operational plant data, and translating successful approaches into software that can eventually run continuously in industrial environments. You will work closely with the founder, research collaborators, and industrial partners, and have significant influence over KIIO's technical direction and product architecture. Typical problems include: modelling complex and partially observed industrial dynamical systems; detecting subtle process deviations under changing operating conditions; localising faults and estimating their severity; distinguishing sensor, actuator, and process-related faults; reasoning about fault propagation through interconnected processes and control loops; building reliable evaluation and deployment pipelines for real-world industrial data. Qualifications We are looking for someone with strong technical depth and, equally importantly, the desire to turn research into something that works in the real world. Master's or PhD in Machine Learning, Computer Science, Control, Signal Processing, Applied Mathematics, Physics, Mechanical/Chemical Engineering, or a related quantitative field. Strong background in one or more of: fault detection and diagnosis, dynamical systems, time-series modelling, system identification, graph machine learning, signal processing, or control. Strong mathematical and algorithmic reasoning, with the ability to formulate ambiguous real-world problems rigorously. Experience developing and evaluating machine-learning or data-driven methods on real datasets. Strong programming skills, particularly in Python and modern scientific/ML tooling. Ability to write clean, maintainable software and translate research prototypes into robust implementations. Comfortable working with noisy, incomplete, high-dimensional sensor data rather than curated benchmark datasets. Interested in understanding why a system behaves the way it does, rather than treating models purely as black boxes. Comfortable working closely with process engineers and industrial partners to translate domain knowledge into technical solutions. Able to work independently and take ownership in a small, fast-moving team. Strong written and spoken English. Particularly relevant backgrounds include PhD researchers or engineers working on fault diagnosis, system identification, process monitoring, industrial AI, time-series ML, graph learning, or data-driven control who want to move from research toward building a real product. #J-18808-Ljbffr
