Deep learning Expert

I'm C., passionate about deep learning, focusing on computer vision tasks like depth estimation, object detection, and scene understanding. Feel FREE to contact me As soon as possible to give you the best ai solutions.

• Computer Vision
• Data Science
• Deeplearning4j4 J.
• flask
• Keras4 J.
• onnx
• Python4 J.
• Pytorch4 J.

During my Ph.D., I proposed a state-of-the-art deep learning model called DepthHist, which achieved an outstanding accuracy of 96.8% on the accuracy (δ1) metric in the widely used benchmarking KITTI dataset. This model, currently under review, uniquely fuses CNNs and Vision attention mechanism with the histogram estimation technique to tackle complex spatial perception challenges from camera data (RGB image) to predict an accurate LiDAR depth map. I also created LiDARDatabox to easily access the most popular LiDAR datasets like KITTI, SUNRGB-D, and Diode with one command. I have open-sourced this tool, and many researchers find data collection easier with it.
Additionally, I have freelanced as an AI engineer, taking up a machine learning project right from conception to deployment. My current skill set ranges from data preprocessing to model building and full-scale system deployment, inclusive of deep proficiency in TensorFlow, PyTorch, and Python.

Technical skills:
1. Environment Setup: Expertise in configuring project environments using Anaconda, Virtual Environments (venv), and Docker, ensuring seamless development and deployment workflows.
2. Python and AI Tools: Proficient in utilizing Numpy, Matplotlib, TensorFlow, and PyTorch, demonstrating comprehensive expertise in model development.
3. Web development: Familiarity with Flask, SQL, HTML, JS, and CSS for web application development.
4. Model Development: Proficient in creating and implementing advanced deep learning models, using Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Transformers.
5. Computer Vision Expertise: Specialized in computer vision applications with a focus on classification, object detection, depth estimation, and image segmentation.
6. Deployment Efficiency: Capable of converting models to TensorFlow Lite (TFLite) and Open Neural Network Exchange (ONNX) formats, ensuring efficient deployment of machine learning models.