Researcher

RISE Lab, Arizona State University

• Engineered automated ROS + Python data acquisition pipeline synchronizing high-frequency pneumatic pressure streams (100+ Hz) with motion-capture trajectories, eliminating manual preprocessing and accelerating analysis workflows by 60% for downstream control modeling.
• Decoded complex system dynamics through MATLAB/Python signal analysis, uncovering critical temporal dependencies and non-linear transients that directly shaped feature engineering strategies for closed-loop pneumatic controllers.
• Spearheaded baseline control model development using regularized linear regression with GridSearchCV hyperparameter optimization, achieving 7% improvement in predictive accuracy over existing approaches and establishing performance benchmarks for the research team.
• Orchestrated cross-institutional collaboration with Virginia Tech researchers, pioneering advanced nonlinear control strategies including AutoTS forecasting, LSTM neural controllers, and policy-gradient RL that enhanced system stability by 12% in dynamic pressure scenarios.

Machine Learning Engineer

Jupiter AI Labs (Freelance)

Lawn AI:

• Transformed property assessment workflows by productionizing end-to-end YOLOv8 instance segmentation pipeline that processes Regrid API satellite imagery, deployed as production-grade Flask REST API on AWS with S3 integration, JWT authentication, and comprehensive error handling serving 500+ daily requests.
• Curated and augmented dataset of 1,700+ satellite images with precise boundary annotations, training custom YOLOv8 model that achieved 95% segmentation accuracy and >90% precision in extracting property boundaries—outperforming baseline approaches by 23%.
• Engineered geometric algorithms for square-footage calculation from segmented polygons, delivering <15% mean absolute error that enabled accurate lawn area estimation critical for client pricing models.

• Architected intelligent health assistant by integrating Rasa conversational AI framework with GPT-3.5 for contextual query understanding, handling 1,000+ health-related conversations weekly with 92% user satisfaction ratings.
• Implemented semantic search infrastructure using ChromaDB vector database for clinic recommendations, incorporating RAG (Retrieval-Augmented Generation) with content filtering guardrails that reduced inappropriate responses by 98% while maintaining personalized, context-aware assistance.

Data Science Intern

Incipient Technologies Pvt. Ltd.

• Delivered high-stakes computer vision solutions for confidential client projects, implementing custom object detection models using transfer learning and data augmentation techniques that consistently achieved ~95% accuracy on proprietary datasets while maintaining strict confidentiality protocols.
• Collaborated cross-functionally with web development team to enhance user interfaces, optimizing frontend performance and implementing responsive design patterns that improved site usability scores by 30% based on user testing feedback.

Student Volunteer

ASU International Students & Scholars Center (ISSC)

• Orchestrated logistics for 15+ multicultural events serving 500+ international students, collaborating with volunteer teams across planning phases and day-of execution to ensure seamless event experiences that fostered cross-cultural connections.
• Served as primary point of contact during events, proactively addressing attendee inquiries, disseminating event schedules, and efficiently managing registration workflows that processed 100+ check-ins per event with zero wait-time complaints.
• Optimized event operations by coordinating setup/teardown logistics, managing material distribution, and refining space layouts based on crowd flow observations—improvements that reduced setup time by 25% and enhanced attendee satisfaction ratings.