• 🦉 Barn Owl Technologies - Creating & deploying insect detection models on AWS using Docker
• 🚀 Building production-ready CV systems
• 📊 Helping farmers optimize pesticide usage

• ☁️ AWS Cloud Architecture
• 🐳 Docker Containerization
• 🤖 Advanced Computer Vision

Software Engineering Intern | Managing the whole process of creating and pushing computer vision models to production at a start up company!

• 🔍 Dataset Curation: Curated and annotated a custom bug detection dataset of 5,000+ images, applying data augmentation to improve model performance under varied lighting and weather conditions
• ☁️ AWS Deployment: Independently deployed trained models on AWS (EC2, S3, Docker ECR), enabling 50+ IoT-connected traps across New England to upload images for real-time inference and bug classification
• 🗄️ Data Pipeline: Managed image and metadata pipeline using PostgreSQL; wrote SQL scripts to organize predictions, track trap performance, and power frontend data display
• 🤝 Cross-team Coordination: Coordinated with app developer and ML advisor to ensure updates were aligned without disrupting codebases
• 📋 Note: Also took photos for the product (click the preorder button on the Barn Owl Technologies website to see!)

Full-Stack Developer | Built a full-stack project management app with AI add-ons using OpenAI, React, Firebase, Next.js, Shadcn, and TypeScript

• 🚀 Production Deployment: Deployed the app on Vercel with implemented authentication and storage systems
• 🔧 Data Management: Used React-Firebase-Hooks for efficient real-time data management and state synchronization
• 💰 Grant Achievement: Secured $1,000 from Babson College's Foundry Grant for further development and feature expansion

iOS Developer | Developed a Swift app using Apple's body pose detection to support cancer rehabilitation for adults 65+

• 📱 Swift Development: Built iOS app with Apple's body pose detection framework for rehabilitation exercises
• 👥 User Testing: Piloted at a senior center and iteratively improved based on participant feedback and usability studies
• 🤖 AI Backend: Built Python backend integrating with OpenAI Realtime API to deliver custom workouts like a physical therapist's guidance
• 🎤 Accessibility Focus: Implemented custom voice function calling to enable full app navigation without requiring touch interaction

Research Assistant | Addressed bias in Automatic Speech Recognition (ASR) models against stuttered speech using LibriSpeech/Stutter datasets

• 🔬 Bias Analysis: Identified significant disparities in Word Error Rate (WER) across major ASR models when processing stuttered speech
• 📊 Model Evaluation: OpenAI's Whisper showed a 2x increase in errors, while Facebook's Wav2Vec showed a 6.4x increase
• 🏆 Research Recognition: Helped create research poster for AAAS presentation, receiving an Honorable Mention at the convention
• 🎯 Impact Focus: Contributed to accessibility research highlighting the need for more inclusive AI systems