New York, New York, United States
Currently in TBDS product team.
Wharton Research Data Services (WRDS), Software Engineering Intern 10/2025 – Present - Designed and implemented a real-time distributed audio processing system for clinical use cases using Docker, AWS ECS Fargate, WebSocket streaming, and Redis. Wharton Research Data Services - Built an incremental intelligent audio pipeline (VAD, speaker diarization, ASR, role identification) as modular containerized components, enabling real-time clinician/patient separation and continuous transcription output. - Implemented a patient-scoped Retrieval-Augmented Generation (RAG) system using ChromaDB and GPT APIs, enforcing data isolation via patient IDs and ensuring upserting data writes across distributed services. - Developed a web-based user interface using React.js, built with Vite 5, leveraging modern JavaScript (ES6) and HTML/CSS for routing, state-driven views, and client-side navigation. - Acted as an AI-native engineer, leveraging Cursor, Kiro, and AI-assisted coding tools throughout system design, containerized service development, and rapid iteration, significantly reducing the prototype-to-deployment cycle.
Designed and implemented key features for the Remote Sensing Platform, leveraging Python, Pandas, NumPy, and Scikit-learn for data analysis, processing, and integration. Developed and optimized backend systems using Python Flask, creating robust RESTful APIs for large-scale remote sensing data pipelines. Applied Random Forest algorithm and other machine learning models to enhance Leaf Area Index (LAI) computation, achieving a 50% improvement in accuracy, and deployed the models using Docker for scalable production environments. Automated workflows for LAI data fusion, distribution, and monitoring by developing scripts in Python, integrating with Linux shell scripting and scheduling with cron jobs. Collaborated with the technical team to deploy applications on AWS EC2 instances, utilizing S3 for storage and Lambda functions for event-driven processes. Conducted research on cloud detection and removal algorithms in the remote sensing field. Authored testing reports for the cloud detection model using UNet with synthetic cloud to evaluate algorithm performance and advantages.
● Spearheaded the development of the Visual Place Recognition (VPR) algorithm tailored for indoor spaces, leveraging Line Segment Detection (LSD) to bolster precision and reliability. ● Executed the line detection algorithm to achieve robust feature extraction, implementing with Python and OpenCV. ● Employed K-means clustering to partition the database into multiple clusters, comparing different image clusters. ● Utilized TensorFlow to train neural network models on proprietary datasets, aimed at enhancing navigation and analysis capabilities within intricate indoor layouts.
● Engaged in the maintenance of OLT (Optical Line Terminal) products, with a focus on core software management modules like PON (Passive Optical Network) and ONU (Optical Network Unit) ● Led on-site debugging efforts for OLT products, employing software positioning commands in C++ ● Composed a comprehensive Debugging Guide for OLT products and created an internal training video offering detailed instructions for various debugging scenarios, garnering a view count of 400 in the Transport domain ● Developed a comprehensive Project Management System utilizing a microservice architecture. Built the backend service as a web service using Spring Boot, MVC, and Spring Data. ● Implemented DTO entities for data persistence with JPA at the data layer, leveraging MySQL and H2 databases. Utilized RabbitMQ as a message broker to decouple backend services effectively. ● Created a customized exception handler service that logs and monitors exceptions, and sends the data back to a logging service. Developed an authentication manager to screen incoming requests using JWT. ● Designed and implemented several RESTful APIs for project CRUD operations on the server-side. ● Led the project through its full lifecycle, including design, prototyping, implementation, testing, deployment, and monitoring. Worked in an agile development environment, maintaining a fast-paced workflow. ● Gained extensive insight into the telecommunications sector, encompassing Nokia product models, network configuration topologies, and the software development lifecycle.