New York, New York, United States
Built production AI platform infrastructure for SmartGlasses, focusing on agentic AI workflows, third-party integrations, voice-interactive experiences, and reusable tooling for real-world deployments. AI Extensibility, New York, NY - Agentic AI Platform: Built foundational WebHook and ClientHook infrastructure enabling third-party applications to integrate with the SmartGlasses AI agent, resulting in a reusable integration layer for external partner workflows. - Chaining AI Queries: Designed and implemented a config-driven JSON transformation en- gine to chain outputs from one external API call into inputs for another, enabling multi-step AI workflows without partner-specific hardcoding. - Function Registry: Created a centralized registry for AI tool routing, schema configuration, function metadata, validation, and error handling, reducing repeated lookup overhead and allowing new integrations to be added through configuration. - Developer Tooling: Built Wearables Developer Center components for self-serve API map- ping, default value validation, and webhook method/URL configuration, enabling external de- velopers to configure integrations with less platform engineering support.
Built large-scale backend and ML-adjacent data systems focusing on production pipelines, model integration, data quality, cost optimization, and operational reliability. Amazon Photos, Seattle, WA (Nov 2021 - Feb 2024) - Image Sentiment: Integrated in-house ML models, including image embeddings and face detection signals, to derive sentiment scores for customer photos, enabling ML-powered photo understanding across billions of images. - Sentiment Backfill: Designed and implemented a large-scale backfill pipeline to compute and persist sentiment scores into live production databases, processing billions of photos reliably. - Cost Optimization: Helped deprecate a redundant DynamoDB table and migrate dependent workflows, reducing infrastructure cost while preserving production data correctness. Amazon Business, Seattle, WA (Mar 2021 - Nov 2021) - Data Validation Engine: Built a data validation engine to detect and fix business pipeline discrepancies, improving data quality and system reliability. - Pipeline Health Improvement: Improved pipeline health through operational fixes, inte- gration testing, monitoring, and documentation, reducing downtime and knowledge gaps
Healthcare Software, Madison, WI - Software Modernization: Rewrote legacy healthcare software using .NET and MVVM ar- chitecture, improving maintainability, scalability, and long-term code quality. - Remote Anesthesia Supervision: Built real-time visualization components for anesthesi- ologists to monitor multiple patients under anesthesia simultaneously, including patient data cards, live vitals graphs, trend views, and alert indicators to support timely clinical intervention. - EHR Development: Developed web applications for large-scale patient record workflows and mentored interns through code reviews, stand-ups, technical guidance and onboarding.
Voice Assistant Development: Developed and maintained features for the SVoice Android app, delivering bug fixes and product updates that improved app quality and user experience.