About

Kiran Gude is a skilled Deep Learning Engineer with a strong background in designing and developing Autonomous Driving vision AI stack and optimizing deep learning algorithms. Currently working at Qualcomm, Kiran is responsible for developing Quantization algorithms to optimize deep learning models to run efficiently on NSP hardware, developing Multi-task neural networks which have a single encoder and multiple decoders that can learn multiple tasks simultaneously, and designing vision AI pipeline with multiple camera environments. Before joining Qualcomm, Kiran worked as a Master Thesis Project at BMW, where she implemented 3D Visual SLAM using Intel Realsense D435i RGB-D cameras for navigation and manipulator tasks in production plants. She also developed PyTorch-based C++ inference engine on NVIDIA GPU with TensorRT framework to boost inference speed by 3-5x. Kiran also worked on dense 3D map reconstruction using voxblox and implemented semantic segmentation of 3d map to add spatial information to the navigation stack. She trained DeepLabv3 TensorFlow based semantic segmentation model to annotate the 3D map with semantics. Kiran’s development environment included ROS and gazebo for simulations. Prior to BMW, Kiran worked as a Software Engineer at Cisco, where she developed Embedded Software for Cisco IOS-XE and XR router platforms, developed Linux device drivers and firmware changes according to hardware schematics, and developed software modules for software-defined networking (SDN) based features for Cisco Aggregation routers. She also developed a Python automation framework for above feature test validation. Kiran holds a Master of Science degree in Autonomous Systems from Technische Universität Berlin. She possesses strong technical skills in Deep Learning, AI, Neural Networks, TensorFlow, Pytorch, Mobile, Frontend, Backend, Software Engineering, and iOS. With 6.43 years of relevant experience, Kiran is a highly skilled and experienced Deep Learning Engineer with a proven track record of designing and developing AI-based solutions for autonomous driving and production plants.