Aman Joshi

Software Developer @Focus Research Labs

Project: Autism Detection System (Flask, TensorFlow, OpenCV, Dlib)• Reduced model loading and inference time from 15s to 3s, improving responsiveness by 80% and enabling smoother real-time predictions.• Maintained 93% accuracy during training and validation by applying advanced optimization techniques and hyperparameter tuning.• Designed and deployed a scalable Flask-based backend architecture, supporting seamless model integration with mobile and web interfaces.• Streamlined data preprocessing and feature extraction pipelines, improving pipeline efficiency and maintainability.• Collaborated cross-functionally on backend optimization, ensuring high performance and stability under production-level workloads.• Leveraged Docker for containerized deployment, ensuring portability and reproducibility across environments.