Resume

Amazon Web Services (EC2 Infrastructure Science) · Seattle, WA

• Own production science strategy for AWS Core Compute hardware diagnostics and repair recommendations, translating ambiguous fleet-recovery constraints into ML decision systems that improve yield, reduce unnecessary part replacement, and scale automated remediation.
• Built production multi-source telemetry fusion pipelines for post-failure hardware component diagnostics, aligning system event logs, memory diagnostics, crash reports, lifecycle events, host snapshots, and repair history into component-level recommendations.
• Architected and deployed production ML models for component diagnosis, slot and part localization, and failure clustering, driving production server repairs at fleet scale, including advanced-accelerator hardware.
• Established leakage-aware temporal evaluation and production calibration for noisy repair-outcome data, including label-taxonomy refinement, component-specific confidence thresholds, and routing policies, improving repair success over the prior baseline.
• Built an end-to-end Monte Carlo reliability analysis for EC2 Live Update failover planning, modeling deployment schedules, notice-policy semantics, and multi-host workload exposure for high-availability customer workloads.
• Designed a production stratified A/B testing framework for causal evaluation of ML routing and remediation changes, and drove roadmap alignment across repair operations, hardware, software, and product.

Amazon (AWS Just Walk Out Technology) · Seattle, WA

• Led capacitive sensing science for AWS Just Walk Out and served as senior scientist on RFID-enabled JWO, designing multimodal event-detection algorithms across capacitive shelf sensors and RFID retail and stadium deployments.
• Built a thin capacitive shelf-sensor pipeline for pick, return, and no-action detection, fusing multi-channel time series into per-lane product hypotheses with channel-to-lane fusion, CUSUM change detection, and Bayesian product-capacitance learning.
• Improved RFID event detection through antenna and algorithm work, and introduced an overlapping-area figure of merit (OA-FOM) for antenna discrimination, with a Python tool that hardware engineers used to evaluate antenna configurations against in-field data.
• Applied attention-based transformers to model RFID tag transitions and built automated root-cause analysis pipelines that reduced algorithm analysis effort.
• Pioneered RFID-enabled Just Walk Out for NFL and MLB stadium retail, and drove science roadmap alignment across research, hardware, firmware, and ML stakeholders on lane mapping, confidence scoring, environmental compensation, and cross-talk.
• Served as Amazon research liaison to Professor Joshua R. Smith's Sensor Systems Lab at the University of Washington (with PhD student Kedi Yan) on wireless power transfer for retail and robotic applications.

Northrop Grumman Mission Systems · Baltimore, MD

• Engineered and optimized a wideband phased array antenna system for a Navy EA-18G program as RF technology lead, validating performance across 150k+ data points.
• Applied machine learning and computer vision to automatic target recognition (ATR) on synthetic aperture radar (SAR) imagery, contributing RF systems expertise across modeling, analysis, and design optimization.
• Led model-based systems engineering and authored proposal sections of a successful $33M engineering contract. Promoted twice: RF Engineer to Principal (2018) and Senior Principal (2020).
• Co-founded InventNG, an internal innovation group, and ran deep-learning and additive-manufacturing hackathons that seeded roughly $100k in follow-on innovation funding.