Designing resilient intelligent systems for conservation, healthcare, and consumer apps.

Delivered Root Cause Analyzer, a full-stack macOS AI application surfacing ticketing insights.

• Achieved 80% top-10 accuracy on 10K+ change tickets, cutting manual triage from about four hours to under 30 minutes.
• Developed the end-to-end SwiftUI, Combine, Foundation, and SwiftData experience from initial research through internal deployment.
• Boosted prediction precision by 15% via Stash API pull-request analysis across the top 10–50 root-cause candidates.
• Shipped Code Fix Analysis recommendations that suggest remediation strategies for approved root causes.
• Implemented hybrid local/API caching to keep historical investigations instantly accessible for cross-functional teams.
• Built React, PostgreSQL, and Flask dashboards charting bug classes and change volumes for the Communications App organization.
• Automated build telemetry ingestion with Kubernetes jobs that tap internal build tracking APIs.

Leading bioacoustic intelligence efforts with the San Diego Zoo Wildlife Alliance and UC San Diego partners.

• Manage and mentor a 15-student team spanning machine learning, software engineering, and data science tracks.
• Deploy CNN and Transformer pipelines that classify Amazon rainforest bird calls alongside SDZWA collaborators.
• Engineer PyTorch, Rust, and ONNX runtimes that deliver over 10× inference and training speedups for large acoustic models.
• Ship Electron.js desktop tooling that unifies labeling workflows, inference jobs, and MySQL-backed data storage.
• Develop embedded analytics on STM32 microcontrollers for low-power wildlife monitoring at KRG.
• Coordinate a 10-person tinyML pod leveraging Kubernetes and AWS Batch to process thousands of audio hours with >95% accuracy — see Website, Report, Paper, and Article.

Advanced a nationwide sleep study measuring incentive programs versus rest outcomes for 30K+ college participants.

• Led cross-platform app development in React and Flutter serving 100+ U.S. campuses.
• Integrated Fitbit, GCC, and Firebase services so longitudinal sleep data flowed seamlessly into analytics pipelines.
• Ran NLP research on PPP policy datasets using NumPy, Pandas, TensorFlow, PyTorch, and Scikit atop SDSC's Narrows cluster.

Scaled internal IoT analytics for automotive partners with Azure-first .NET services.

• Built serverless image-processing pipelines with .NET, Microsoft Azure, C#, and FFmpeg to auto-generate daily timelapse analytics.
• Created Azure DevOps workflows that retrieve, process, and catalog imagery and metadata from REST APIs.
• Leveraged Azure Blob Storage and Durable Functions to productionize internal tooling end-to-end.

Led multi-camera vehicle scanning systems spanning Mercedes-Benz, Porsche, and Amazon logistics operations.

• Directed a five-engineer team building damage-detection platforms that automatically classify vehicle imagery (Website).
• Re-trained VGG-19 models with 50TB+ of telemetry on custom supercomputers, surpassing 90% accuracy for Mercedes-Benz and Amazon warehouses (Github).
• Developed a Flutter-based Porsche internal App Store experience to locate scans under logistics leadership mentorship (Letter).
• Architected low-latency REST APIs and streaming data pipelines ingesting 2TB+ of daily telemetry from 20+ European scanners.