I've co-led AeroML at Pratt & Whitney Canada, a no-code AutoML platform for aerospace manufacturing that achieved 38.2% lower MAE than the company’s best internal model and reduced advanced model development from weeks to minutes. I built its local-first Python/FastAPI platform with secure on-device processing and optimized a 955 MB dataset from 10.4 seconds to 4.35 seconds using DuckDB and PyArrow.
At Ericsson, I built and debugged backend services and automation tools for dependency, vulnerability, and compliance workflows. I also lead end-to-end development of Secret Garden TV’s upcoming digital platform, covering requirements, architecture, backend services, cloud infrastructure, CI/CD, and long-term maintainability. My projects include Asyncra, a distributed job orchestration system benchmarked at 874 jobs per second across 16 workers, and AtlasML, a reproducible ML execution platform with explicit service contracts.
