Currently an Extern @Wayfair
I build end-to-end apps that combine ML models, real-time computer vision, and cloud services to turn research into working products.
Work includes real-time CV pipelines, AutoML-integrated therapy tools, and voice-first wellness apps, built with FastAPI, Next.js, OpenCV, AutoML, and cloud storage.
Built a full-stack AI soccer juggling coaching tool (Python FastAPI, OpenCV, NumPy, React/Vite) deployed on Kaggle GPU, demonstrating end-to-end ML model deployment in a cloud-native environment. Engineered a real-time CV pipeline combining YOLO detection, Kalman filter smoothing, and MediaPipe pose estimation to classify actions and compute statistical rhythm metrics.
Developed a Next.js digital therapy platform for Parkinson s patients, integrating Google Vertex AutoML and custom scoring for multimodal handwriting analysis. Built a spiral calibration pipeline on the platform that combines AutoML inference with statistical scoring to produce a 0 0100 tremor-severity score from patient drawing data, with images and assets stored in Amazon S3.
I build end-to-end apps that combine ML models, real-time computer vision, and cloud services to turn research into working products.
I am Pranav Gunjala, a junior Computer Science student at the University of Florida and a Full-Stack Developer on Project FUNLIT. I have backend internship experience, worked on undergraduate ML research, and build full-stack AI and computer vision projects that combine models, pipelines, and cloud deployment.
Externships
Wayfair n8n AI Agent Engineering Externship
Wayfair
Experience
Back End Developer Intern Remote
Florida Community Innovation Foundation · Oct. 2024 – Sep. 2025
Undergraduate ML Research Assistant — Project Align
Virtual Learning Lab — University of Florida · Jan. 2025 – Apr. 2025
Full-Stack Developer — Project FUNLIT
Virtual Learning Lab — University of Florida · May 2025 – Dec. 2025
Education
University of Florida
Bachelor of Science in Computer Science
Skills
Built a full-stack AI soccer juggling coaching tool (Python FastAPI, OpenCV, NumPy, React/Vite) deployed on Kaggle GPU, demonstrating end-to-end ML model deployment in a cloud-native environment. Engineered a real-time CV pipeline combining YOLO detection, Kalman filter smoothing, and MediaPipe pose estimation to classify actions and compute statistical rhythm metrics.
Developed a Next.js digital therapy platform for Parkinson s patients, integrating Google Vertex AutoML and custom scoring for multimodal handwriting analysis. Built a spiral calibration pipeline on the platform that combines AutoML inference with statistical scoring to produce a 0 0100 tremor-severity score from patient drawing data, with images and assets stored in Amazon S3.
Built a full-stack mental wellness platform (Next.js, Go) with journaling at the center: voice-to-text, searchable history, and review, integrated with a Vapi AI voice assistant for conversational, hands-free input and therapy-style support. Fed journal transcriptions into Gemini to recommend therapy sessions and corresponding events in Google Calendar, enabling agentic scheduling grounded in what users wrote or spoke.