Krishnanjana Reddy

• Developed, trained and tested deep learning models for autonomous vehicle tasks with camera and lidar
data using Pytorch.
• Worked with object detection, panoptic segmentation, monocular depth estimation and traffic light
classification tasks.
• Implemented mid-fusion techniques for camera-LIDAR data fusion in BEV space and late-fusion methods lever￾aging objects geometric and semantic consistencies using neural networks.
• Developed tools to automate both the data collection and in-house annotation processes, facilitating the efficient
generation of datasets for training deep learning models.
• Developed custom Convolutional Neural Networks (CNNs) from the initial design phase through to
deployment, incorporating in-house data to meet specific project requirements.
• Worked with training pipelines in SageMaker for training new batches of data at scale and monitoring perfor￾mance.
• Optimised deep learning models for performance using TensorRT, proficient in converting advanced ar￾chitectures, implemented custom operators, and wrote CUDA modules for sensor data processing.
• Participated in multiple autonomous vehicle tests aimed at identifying the limitations of the current stack within a
specific operational design domain (ODD) and collaborated with other teams to develop safer solutions.
• Conducted thorough research and testing of state-of-the-art sensing systems (cameras and lidars) and
perception algorithms for the next-generation autonomous vehicle stack, creating detailed comparison reports.
• Showcased autonomous vehicle operations to clients from both industry and academia.
• Played a key role in upgrading the autonomous vehicle stack from ROS 1 to ROS 2.
• Customized the sensor drivers at a lower level to ensure that they provide sensor data and diagnostics in
accordance with our AV stack.

<p>Developed, trained and tested deep learning models for autonomous vehicle tasks with camera and lidar data using Pytorch.<br>
Worked with object detection, panoptic segmentation, monocular depth estimation and traffic light classification tasks.<br>
Implemented mid-fusion techniques for camera-LIDAR data fusion in BEV space and late-fusion methods leveraging objects geometric and semantic consistencies using neural networks.<br>
Developed tools to automate both the data collection and in-house annotation processes, facilitating the efficient generation of datasets for training deep learning models.<br>
Developed custom Convolutional Neural Networks (CNNs) from the initial design phase through to deployment, incorporating in-house data to meet specific project requirements.<br>
Worked with training pipelines in SageMaker for training new batches of data at scale and monitoring performance.<br>
Optimised deep learning models for performance using TensorRT, proficient in converting advanced architectures, implemented custom operators, and wrote CUDA modules for sensor data processing.<br>
Participated in multiple autonomous vehicle tests aimed at identifying the limitations of the current stack within a specific operational design domain (ODD) and collaborated with other teams to develop safer solutions.<br>
Conducted thorough research and testing of state-of-the-art sensing systems (cameras and lidars) and perception algorithms for the next-generation autonomous vehicle stack, creating detailed comparison reports.<br>
Showcased autonomous vehicle operations to clients from both industry and academia.<br>
Played a key role in upgrading the autonomous vehicle stack from ROS 1 to ROS 2.<br>
Customized the sensor drivers at a lower level to ensure that they provide sensor data and diagnostics in accordance with our AV stack.<br>
Utilizing various open-source perception algorithms, we enhanced their speed and accuracy by tailoring them to specific task requirements and incorporating the latest advances in computer vision and machine learning.<br>
An end-to-end scalable pipeline that facilitates nearly continuous data collection, annotation, training, and testing of deep learning models.<br>
This experience comprises a two-year remote job based in India. My performance was so impressive that I was offered the opportunity to work closely with an Autonomous Vehicle team in the UK, allowing me to contribute more effectively.</p>

<p>Conducted in-depth analysis of AI technical literature, including patents and research papers.<br>
Derived valuable insights on AI’s impact on client industries.<br>
Provided clients with competitive research directions and facilitated IP management for strategic advantage.<br>
During his tenure with us, we found him to be hardworking and very productive.<br>
We found him to be a self-starter who is motivated, duty bound and a highly committed team player with strong conceptual knowledge.</p>