Experience
I build machine learning systems end to end: architecture, training data, experimentation, orchestration, evaluation, and deployment. Client-facing work below is described in generalized terms to avoid exposing unnecessary implementation details.
Larus Technologies
Data Scientist
2023 - Present
Build and deploy applied AI systems across enterprise, aviation, defence-adjacent, and sensing programs.
- Architected hierarchical NLP classification and active-learning workflows for retail taxonomy modernization spanning more than 400M product records.
- Built RAG pipelines and evaluation tooling for grounded report generation and briefing support.
- Developed Databricks + MLflow forecasting, pricing, and anomaly-detection workflows for aviation aftermarket decision support.
- Delivered multimodal sensing and remote-sensing pipelines using EO, IR, GPR, magnetometer, and satellite imagery.
Stack: Python, PyTorch, Transformers, Databricks, MLflow, PySpark, Kubernetes, FastAPI
Outlier AI
Contributor and Reviewer
Selected engagements
Evaluated frontier AI models in mathematics and physics using research-grade prompts, rubrics, and failure analysis.
- Created Ph.D.-level evaluation prompts to probe advanced reasoning and edge cases.
- Designed scoring rubrics focused on correctness, depth of reasoning, and robustness.
Stack: Evaluation design, mathematics, physics, reasoning analysis
University of Ottawa
Ph.D. Researcher, Earth Science
2019 - 2023
Built open-source ML tooling for earthquake detection, denoising, phase picking, and seismic dataset creation.
- Created QuakeLabeler for seismic annotation and dataset generation.
- Built Blockly Earthquake Transformer for configurable seismic model training.
- Published research on deep learning tools for earthquake detection and seismic noise removal.
Stack: PyTorch, seismic ML, signal processing, tooling, open-source research software