Responsibilities

- Apply machine learning, deep learning, and AI techniques to make analysis of multi-sensor infrastructure data more efficient, including 3D point clouds, subsurface radar, imagery, and satellite deformation data
- Build and tune advanced algorithms, regression, clustering, decision trees, neural networks, simulation, and scenario analysis to detect track geometry exceptions, subsurface anomalies, surface defects, and vegetation encroachment
- Develop AI models and deploy them into the company's onboard monitoring units and cloud analytics platform
- Build predictive models that forecast asset degradation and remaining service life, producing risk-ranked maintenance recommendations
- Run large-scale experiments to uncover hidden relationships between ground deformation, climate, freeze-thaw history, and observed infrastructure degradation
- Implement recent statistical and mathematical methods to solve specific business problems, including sensor fusion, change detection across repeat surveys, and data reduction of high-volume sensor streams
- Prepare and extract data using programming languages from cloud systems, edge devices, geospatial databases, and third-party satellite, weather, and hydrological feeds
- Coordinate analysis across structured and unstructured data, using code to clean, georeference, align, and organize it
- Validate model outputs against field campaigns, manual inspection records,



and regulatory condition thresholds, and quantify accuracy and uncertainty
- Visualize defects, deformation trends, and geospatial risk layers so clients and internal teams can quickly conclude
- Present findings to railway engineering executives, maintenance planners, and government partners, and support pilots, RFP responses, and collaborative R&D; projects

Qualifications

- Degree in statistics, mathematics, computer science, engineering, geomatics, or a related quantitative field; master's or doctoral degree preferred
- 2+ years applying statistical modelling or machine learning to real-world, noisy data
- Strong programming skills in Python and SQL
- Hands-on experience with at least one sensor-derived data type: point clouds, radar, imagery, time series, or remote-sensing rasters
- Solid grounding in experimental design, model validation, and uncertainty quantification
- Clear written and verbal communication with non-specialist audiences

Nice to Have

- Geospatial data and GIS tooling; coordinate systems and georeferencing
- Computer vision, 3D deep learning, signal processing, or SAR/InSAR processing
- Cloud or edge model deployment and MLOps practices
- Railway, civil, geotechnical, or asset-management domain exposure
- Willingness to travel occasionally for field data collection, including remote and northern sites

#J-18808-Ljbffr

📌 Lead Data Scientist (Toronto)
🏢 ApoSys Technologies
📍 Toronto