Advancing applied intelligence for manufacturing, autonomous vehicles, and public-sector infrastructure.

Manufacturers face a persistent trade-off between production speed and quality. Traditional workflows rely on reactive maintenance and manual inspection, resulting in costly downtime and defect-driven losses. I built cloud and AI prototypes that unify vision-based defect detection, IoT-driven predictive maintenance, and digital twin simulation into a single integrated manufacturing intelligence platform on AWS.

A representative engagement is the real-time customer diagnostic solution I architected for Sonos on Amazon OpenSearch Service, which reduced their operational overhead by approximately 90 percent with zero outages since migration. In luxury automotive manufacturing with Lucid Motors, vision-based quality inspection catches misaligned assemblies before they disrupt downstream airflow systems, where a single defect can cost thousands per vehicle. My co-authored AWS GenAI Atlas chapter on Case Studies for Manufacturing documents these patterns, and my IEEE ICCIDS 2026 paper reports field results of up to 42 percent defect rate reduction, 38 percent reduction in unplanned downtime, and 23 percent improvement in overall equipment effectiveness.

Unplanned downtime costs US manufacturers an average of 260,000 dollars per hour, adding up to roughly 50 billion dollars per year. By making AI-driven manufacturing practical for mid-sized manufacturers and not only large enterprises, this work advances US reshoring economics and directly addresses one of industry's most expensive problems.

Advanced Driver Assistance Systems span Level 1 through Level 4 autonomy, from adaptive cruise and automatic emergency braking to fully autonomous driving. Test fleets generate terabytes of sensor data daily from cameras, LiDAR, and radar, all of which must be ingested securely, labeled at scale, and used to train models that make safety-critical real-time decisions.

I architect end-to-end solutions on AWS's Autonomous Driving Data Framework that cover sensor ingestion, cloud transmission, labeling, and model training. Where certain edge cases cannot be safely or economically captured in the real world, such as pedestrians entering a roadway at night or obstacles in adverse weather, I use synthetic data generation to complete training coverage. My work with Lucid Motors applies this at production scale, supporting their ADAS development pipeline.

Beyond infrastructure, I contribute to the reliability of AI in safety-critical automotive contexts. My co-authored AWS GenAI Atlas chapters on Deploying Generative AI Applications and Managing Hallucinations and Guardrails address the validation and guardrail patterns needed when a false positive can mean the difference between safe operation and a collision. My features in Towards Data Science, Hyperight, and Packt's DataPro newsletter extend this thinking to how generative AI will reshape vehicle design itself, the economics of synthetic data, and regulatory readiness for explainable automotive ML.

Government networks carry sensitive citizen data and critical national infrastructure that must be modernized without compromising sovereignty or security. As Senior Solutions Architect at TELUS, I architected secure cloud infrastructure for the Government of Canada and the Government of Ontario under TELUS's 176 million dollar federal modernization contract, building deployment pipelines entirely in Terraform to enable private-sector velocity with public-sector audit, compliance, and control.

In my final year at TELUS, I was a core architect on the company's launch of Canada's leading sovereign AI factory powered by NVIDIA, piecing together the IT service management architecture that supports domestic AI capabilities without foreign infrastructure dependence. The work brings together data sovereignty guarantees, encryption and access controls, and operational resilience patterns drawn from safety-critical manufacturing and automotive domains, applied to government-grade workloads.

The outcome is measurable: services that previously took weeks to deploy can now be updated in hours, government-grade security posture is preserved end to end, and Canada gains strategic independence for AI workloads that previously depended on non-domestic infrastructure.