B Ravi Kiran

B Ravi Kiran

Multi-modal perception, Staff Engineer, Qualcomm since Nov 2022.

  • 5+ years Deep learning for Camera perception pipeline, Large scale pointcloud processing
  • 12 years working on computer vision, ML/DL applications with broad exposure to different application domains
  • Email : (beedotkiran AT gmail DOT com)
  • LinkedIn, Twitter, Scholar
  • CVs 1page, EN, FR

Experience

Academia (7 years) Industry (6 years)
Oct’16-Aug’17 : ATER Université Lille 3 & CRISTAL, Nov’18-Oct’22 : ML/DL for perception Technical Lead, Navya
Dec’15-Nov’16 : Post-doc DATA-ENS Paris & ThalesAlenia Space, Jan’18-Oct’18 : R&D Engineer, Autonomous Systems, AKKA Technologies
Nov’14-Nov’15 : Post-Doc CAOR Mines ParisTech & Helicoid May’17-Dec’17 : Vision & deep learning consultant Uncanny Vision
Oct’11-Oct’14 : PhD A3SI-LIGM, Univ. Paris-Est Aug’08-Feb’10 : Embedded Software Engineer, Texas Instruments, Bangalore, India

Current Domains

Computer vision/Cameras 3D-DL/LiDARs
Deployment pipeline for realtime Object detection Joint work with Thomas Gauthier Active learning for large scale semantic segmentation Joint work with Anh Duong, Alexandre Almin, Leo Lemarie
Realtime Multi-task learning architectures for Autonomous Driving Joint work with Hao Liu & Thomas Gauthier 3D-Deep learning for Large scale segmentation pointclouds, domain adaptation & data augmentatations : For Semantic Maps Joint work with Fouad Makiyeh, Frederic Ferre, Alexandre Almin Nicole Camous **
Navya Perception Datasets (N3DS & NODA) Joint work with ML team Alexandre Almin, Hao Liu, Leo Lemarie Thomas Gauthier Simulation-to-real domain adapatation for 3D object detection in Pointclouds : to build offboard large scale 3D annotator on pointclouds Joint work with Weishuang Zhang, Thomas Gauthier, Yanis Mazouz, Theo Steger
SSL and geometry preserving DA for Monocular 3D object detectors pdf LiDAR for Autonomous Driving literature survey
  • Focus on Deep/Reinforcement learning methods for Autonomous driving tasks, supervised PhD & Master's internships on the topic
  • Building better understanding of perception-planning interface in AD pipeline

Recent work

  • Navya3DSeg Dataset paper Feb 2023, presentation at 3D-DLAD IV Aug 2022 talk
  • Point cloud representations for deep learning on Pointclouds, course presentation ENSTA Paris 2022
  • Real world data augmentations for perception tasks in autonomous driving Speaker Auto.AI 2021 Berlin, Slides
  • Exploring real-world reinforcement learning for Autonomous Driving Speaker Cognitive Vehicles 2019 Slides

Deep Learning for AD Workshops Series YouTube Channel

Reviewer

Transaction-ITS 2020-22, PR & PRL 2013-21, ITSC 2016-2022, ISMM 2013-19, ML4AD 2019-2022, ICRA-AV workshop 2022, IROS 2017, TIP 2017, ICVES 2017, ICIP 2014,