Robot with 360-degree vision system using SLAM technology for navigation, featuring advanced path planning algorithms and sensor integration.
Utilized CNNs, Random Forest, and SVM for satellite image classification. Implemented Sci-kit learn toolkit and Anaconda environments. Used HPC clusters on Linux with NumPy, Pandas, Matplotlib, and Plotly.
Collaborated on all Dyson's air purifiers embedded software, designed backbone logic behavior system for hardware device communication, and updated fundamental library stack for 10+ projects.
Implemented CNNs for computer vision tasks including image classification, object detection, and segmentation. Optimized architectures using Transfer Learning and TensorRT for autonomous vehicle applications.
Android application for edge ML model deployment using YOLO, Mask R-CNN, and SSD frameworks, optimized with TensorFlow Lite for efficient inference.
Real-time object detection and tracking system using state-of-the-art deep learning models and efficient inference techniques.
Implementation of a large language model training pipeline with efficient attention mechanisms and distributed training support.
End-to-end platform for deploying, monitoring, and scaling machine learning models in production environments.
Comprehensive natural language processing pipeline with support for various tasks including text classification, named entity recognition, and sentiment analysis.
A flexible framework for implementing and experimenting with various reinforcement learning algorithms and environments.
