Projects

Built a RAG agent that guides teams through achieving compliance certifications such as SOC 2, answering privacy and controls questions against current regulatory text. Served on Google Cloud Run.
Ingested regulations, standards, and articles from the web on a daily schedule through Cloud Run jobs, so answers track requirements as they change rather than a stale snapshot.
Embedded the corpus into a vector database for semantic search and served the model through Vertex AI, keeping ingestion, retrieval, and inference on a single platform.

Built the on-device agent in Go for a US security startup's platform, monitoring managed devices across customer fleets and applying control actions to them.
Implemented the device-to-backend channel over a message broker, streaming telemetry upstream and executing commands pushed back down to each device.
Packaged and deployed the supporting services with Docker and Kubernetes.

Designed and developed a Django-based application to appraise farms; platform launched across British Columbia, Alberta, and Saskatchewan.
Integrated heterogeneous data sources — Canadian official sales records, soil information, and satellite imagery — to predict per-farm market prices.
Orchestrated data ingestion and cleaning pipelines with Airflow, providing scheduled runs, retries, and lineage visibility.
Built an internal Django Admin console alongside the user panel so product managers could review user queries, feedback, and underlying data health.
Deployed on AWS, combining EC2 instances with Lambda functions.