Senior ML Data Scientist - Wireless People Sensing

Algorized, a venture‑backed deep‑technology firm rooted in Switzerland and operating from Silicon Valley, creates edge‑AI solutions that let robots perceive humans through existing wireless sensors. The opening is for a Senior ML Data Scientist focused on wireless people‑sensing, based on‑site in Etoy, Switzerland, with no remote option. In this role you will craft and refine machine‑learning pipelines that interpret radar and other wireless signals to deliver reliable people‑detection capabilities. You will build core models that serve a variety of sensing tasks, environments and hardware setups, and devise signal‑processing and deep‑learning techniques that turn raw sensor streams into accurate outputs. Designing model structures that capture spatial and temporal dynamics, setting up experiments, defining evaluation criteria, and analysing performance limits are key parts of the job. You will also stay current with advances in time‑series and representation learning, integrate them to boost robustness, work closely with software engineers and MLOps to build the supporting infrastructure, and guide junior colleagues while shaping the team's data‑science practices. The position requires a doctorate in a quantitative discipline such as data science, computer science, wireless communications, electrical engineering, applied mathematics, physics or a closely related field. Candidates must have at least three years of practical experience delivering machine‑learning or data‑science products for real‑world use cases. A solid track record of applying statistical and learning methods to sensor or time‑series data is essential, as is deep knowledge of signal‑processing techniques like spectral analysis, filtering, detection, tracking or sensor fusion. Mastery of core machine‑learning concepts, including model creation, evaluation, optimisation and generalisation, is expected, together with strong Python programming skills and hands‑on use of libraries such as PyTorch, scikit‑learn, NumPy and SciPy. The ability to own ambiguous technical challenges and drive them from concept to validated solution, plus excellent teamwork and communication abilities, are also mandatory. Additional assets include experience with radar, RF sensing, LiDAR, computer‑vision or acoustic modalities, and a history of designing experiments that cope with noisy, imperfect data. Familiarity with self‑supervised, representation or multimodal foundation‑model approaches, as well as work on 3D positioning, occupancy detection, activity recognition, tracking or vital‑sign estimation, will be viewed favorably. Understanding the constraints of edge‑AI deployment and having previously moved models from research to production are beneficial, as is prior collaboration across data‑science, embedded, software and product groups. What the role asks for: - PhD in Data Science, Computer Science, Wireless Communication, EE, Applied Math, Physics or related field - Minimum 3 years hands‑on ML or data‑science experience for real‑world applications - Proven ability to apply statistical and learning methods to sensor or time‑series data - Strong knowledge of signal‑processing techniques such as spectral analysis, filtering, detection, tracking or sensor fusion - Deep understanding of machine‑learning fundamentals including model development, evaluation and optimisation - Proficient Python programming with PyTorch, scikit‑learn, NumPy and SciPy - Demonstrated ownership of open‑ended technical problems from exploration to validated solution - Excellent collaboration and communication skills for cross‑functional work - Experience with radar, RF sensing, LiDAR, computer vision or acoustics (nice‑to‑have) - Background designing experiments with noisy, imperfect real‑world data (nice‑to‑have) - Familiarity with self‑supervised, representation or multimodal foundation‑model techniques (nice‑to‑have) - Experience in 3D positioning, occupancy sensing, activity recognition or vital‑sign estimation (nice‑to‑have) - Knowledge of edge‑AI constraints and production deployment trade‑offs (nice‑to‑have) - History of collaboration across data‑science, embedded, software and product teams (nice‑to‑have)