Associate Ai/Ml Infrastructure Engineer
Designed and deployed advanced deep learning models for semantic segmentation tasks, specifically road extraction tasks using PyTorch and TensorFlow, leveraging CUDA to optimize model training and inference on GPU environments.Tested and fine-tuned state-of-the-art models, including DPT, Segformer, and SETR, achieving a significant improvement in road segmentation accuracy from remote sensing imagery.Built a scalable pipeline for preprocessing and managing large geospatial datasets, ensuring compatibility with high-performance AI workflows.Collaborated with cross-functional teams to refine model performance, using Git for version control and maintaining a well-documented codebase for reproducibility.