• Supported enterprise PCIe SSD validation and debug, partnering with Application Engineering and system engineers to reproduce failures, isolate root causes, and improve product reliability

• Developed and maintained Python-based test automation to streamline validation workflows, including test orchestration, log collection, and result reporting for system-level testing

• Built AI-assisted engineering utilities using Retrieval-Augmented Generation (RAG) and related workflows to accelerate troubleshooting, knowledge retrieval, and validation triage

• Applied ML techniques for log mining and anomaly detection to surface failure signatures, reduce manual log review, and improve debug efficiency

• Developed an enterprise order management platform with a microservices architecture, utilizing Spring Cloud and Eureka for service discovery, integrated with Feign for remote calls, and employed Spring Data JPA for efficient data persistence and query operations.

• Led optimizing order query response time by analyzing MySQL slow query logs with pt-query-digest, identifying and improving covering indexes and composite indexes, reducing the response time from 2.5 seconds to 112 milliseconds.

• Implemented a permission management module, with dynamic role and permission display and validation on the front end, and utilized Spring Security with JWT for user authentication, role management, and data encryption on the back end.

• Developed and integrated a log management module using Logback and Spring Boot for standardized logging, and integrated with ELK Stack for centralized log storage, querying, and visualization, supporting real-time monitoring and troubleshooting.

• Established a CI/CD pipeline using Docker and AWS, along with Jenkins for automated builds and deployments, reducing release cycles and enhancing deployment stability and efficiency, ensuring the high availability of the microservices architecture.

• Developed a robust vehicle-infrastructure cooperative 3D object detection model to address communication latency and sensor synchrony, achieving ~2.5% improvement in both 3D AP and BEV AP at 500ms latency compared to the baseline (FFNet) in the AIR-V2X-C dataset, as reported in a 2023 NeurIPS paper

• Developed a multi-modal collaborative perception system for autonomous driving by integrating camera and LiDAR data, achieving an 8% performance improvement compared to the original vision-based system

• Collaborated with scientists and domain experts to shape future research directions and experimental focus

• Guided and debugged assignments for over 30 students, delivering detailed feedback and actionable improvement suggestions

• Applied algorithmic techniques such as divide-and-conquer, dynamic programming, and graph algorithms

• Turned query analytics into performance wins by redesigning indexes and access patterns

• Defined logical and physical service architecture for a large information system; produced a host and configuration matrix with SLOs, capacity plan, rollout timelines, and cost targets

• Partnered with vendors to design operational databases and established SAP EDI integrations with Amazon, Best Buy and Wayfair; added idempotent processing, retries and monitoring

• Built data pipelines for demand clustering and forecasting on 200k+ orders using K-Means and ARIMA; generated warehouse placement and inventory recommendations