Medical Device Innovation

• Led firmware development for Athelas Home, the first FDA-approved in-home complete blood count device, enabling at-home monitoring for Schizophrenia patients

• Optimized optical calibration procedures, improving manufacturing yield by 30% and reducing field issues by 50%

• Implemented TensorFlow serving infrastructure for one-click ML model deployment, enabling new features like platelet detection

Infrastructure and Tools

• Developed internal tools suite for real-time manufacturing traceability, remote diagnostics, and over-the-air updates, reducing support response times by 60%

• Architected GCP infrastructure over Kubernetes, cutting compute costs by 40% while supporting high daily test volumes

• Created a comprehensive RPM web portal for real-time patient vital sign monitoring, allowing distributed nurse teams to provide proactive care for over 100k patients

Healthcare Operations and Compliance

• Engineered an automated system for Clozapine REMS reporting in pharmacy operations, ensuring regulatory compliance and reducing manual effort by 80%

• Designed and implemented a remittance ingestion platform processing millions annually across multiple formats for 50+ healthcare practices

• Developed monitoring and alerting systems for full traceability of all transactions in the RCM process, reducing errors by 90%

Process Automation and Efficiency

• Built a task assignment, management, and review system for human-in-loop operations, increasing AR team productivity by 40%

• Reverse-engineered APIs and developed async scraping tools to pull thousands of claims across multiple payers, accelerating claim processing by orders of magnitude

• Created full-stack tooling for the AR team, supercharging manual processes and reducing administrative overhead

Implemented risk management features for Uber's peer-to-peer payments, enabling safe expansion into new financial services while complying with complex regulatory requirements

Transitioned fraud detection from batch processing to real-time risk assessment, significantly reducing financial losses on a platform processing millions of transactions daily

Developed machine learning models around anomaly detection to combat fraud on Uber Eats, addressing the unique challenges of rapid food delivery and protecting both restaurants and Uber's revenue

Created a real-time system to optimize rider pickup locations in challenging urban environments with GPS inaccuracies due to tall buildings and high population density

Utilized big data and clustering algorithms to infer actual rider locations, reducing average pickup times and ride cancellations in dense urban areas

Developed a visualization tool for operations teams, enabling data-driven decisions to improve driver positioning and overall platform efficiency in complex city layouts

Developed firmware for a smart lock system, balancing advanced security features with ultra-low power consumption to achieve one-year battery life on a single CR2 battery

Created on-site debugging tools to rapidly diagnose and resolve issues, crucial for a startup competing against established players in the smart home market