Aarav Rao

Aspiring electrical engineer with hands-on experience in building sustainable tech solutions and co-founding

organizations that promote global AI education. Skilled in Python, computer vision, and AI model development.

Awarded for innovation in STEM competitions and experienced in leading large volunteer teams.

Tennis Instructor @ ACEing Autism

Volunteered as a tennis coach at ACEing Autism, helping children with autism develop motor skills, coordination, and confidence through adaptive tennis instruction.  $594.45 raised through ~20 service hours

RecyclAI - Waste Classification Web App

• Built a full-stack Progressive Web App using React and Flask, leveraging the Google Gemini Vision API to classify waste into 5 categories (Recyclable, Non-Recyclable, Organic, E-Waste, StoreDropOff) with 95%+ accuracy.
• Deployed a scalable inference backend on AWS EC2, enabling real-time classification with confidence scoring and low-latency API responses.
• Integrated a location-based disposal feature that recommends nearby drop-off sites (e.g., e-waste centers and store drop-off locations) to improve proper waste handling and reduce contamination.
• Adopted by sustainability nonprofit One Earth One Chance, supporting broader community education and real-world recycling decision-making.
• recycl-ai.com

SmartBin - AI-Enabled Waste Sorting Hardware

• Designed and built an automated waste-sorting bin integrating camera vision, ultrasonic sensing, and servo-based actuation to classify waste at the point of disposal.
• Leveraged the Google Gemini Vision API to classify items into 5 categories, achieving 90% accuracy across 500+ real-item test trials.
• Achieved an average end-to-end sorting time of 3.5 seconds, enabling near real-time classification and mechanical sorting.
• Reduced sorting error rate by 61% compared to manual sorting, improving recycling accuracy and minimizing contamination.

Appliwaste - Textile Classifying & Recycling Application (In Development)

• Developing the backend for a textile recycling application that classifies fabric composition (e.g., cotton, polyester, linen) from user-uploaded images.
• Integrated the Google Gemini Vision API for material classification using an API-driven inference pipeline with Base64 image encoding.
• Implementing secure authentication and user session management using Auth0, enabling personalized tracking and future analytics.
• Designing scalable REST API endpoints to support real-time classification, recycling guidance, and textile recovery workflows.

Multi-Class Waste Classification Dataset — IEEE DataPort

• Published 100,050 labeled images across 5 waste categories to advance AI research in sustainability.
• Dataset optimized for lightweight CNNs and edge deployment; 941+ views and 61+ downloads in the first year.
• https://ieee-dataport.org/documents/multi-class-waste-classification-image-dataset-cnn-training-and-tflite-deployment-edge

Co-Founder & President of Machine Learning Club

• Founded and lead a student organization teaching core AI/ML concepts (NLPs, LLMs, CNNs) to 50+ students, creating hands-on projects and workshops.
• Developed curriculum and mentored peers to build practical skills in Python, TensorFlow, and computer vision.
• Managed a team of volunteers to support workshops and events, increasing participation by 40% in the first year.

Plano Youth Chapter Lead — One Earth One Chance (OEOC)

• Host Sustainably Wired, a podcast interviewing sustainability and tech leaders, producing 15 episodes with 117+ streams/downloads and 18 Spotify followers.

JV Tennis Team Member 9th & 10th Grade

• Competed on junior varsity team while balancing academic and extracurricular leadership responsibilities.