Gopi Kishan

Experience

• 

  Indian Institute of Technology, Bombay

  Jun 2018 - Jul 2018

  Intern

  I worked on the object detection algorithm, mainly reading the research papers of RCNN, Fast-RCNN, Faster-RCNN and YOLO. I, then built a face detection model using VGG16 in TensorFlow.

• 

  Artificial Intelligence and Electronics Society (ArIES)

  Oct 2018 - Apr 2020

  Project Mentor

  I worked on robustness against Adversarial Attacks. Most Deep learning models are vulnerable to strategically modified samples for a false output with high confidence. I have studied and surveyed various attacks and defence mechanisms with a particular interest in vision models.I have also competed in the Inter-IIT Data Science challenge 2019.Alongside, I have mentored the project "Translational Tools" for SHRISHTI, an Annual Technical Exhibition of IIT Roorkee.

• 

  Indian Institute of Science (IISc)

  May 2019 - Jul 2019

  Intern

  I worked on compression techniques for reducing the size of the deep neural networks. I studied and explored various methods including ​ parameter pruning, weight matrix factorisation, quantisation and knowledge distillation. The compression task was focused on Generative models which inspired me for in-depth knowledge of Variational Autoencoders and GANs.

• 

  ACM Student Chapter IIT Roorkee

  Jul 2019 - Mar 2020

  Moderator

  We organization open discussion on Linux and OS, and mentor projects for campus students.

• 

  Max Planck Institutes Tübingen

  May 2020 - Dec 2020

  Research Intern

  I have developed a physics simulator for microrobotic swarms. For navigation, I applied Deep RL and Imitation learning methods for training agents.

• 

  The University of British Columbia

  May 2020 - Sept 2020

  Visting URA

  Safety properties of a neural network are critical to their application in safety-critical domains. Motivated by this, I worked on verification methods of Deep Learning systems to build safe AI. I have proposed two metrics for probabilistic error quantification of the model hypothesis in my paper titled "Probabilistic Robustness Quantification of Neural Networks" accepted at AAAI-21 Undergraduate Consortium.

• 

  York University

  Jul 2020 - Jul 2020

  CVR-VISTA summer school Attendee

  Awesome talks, lab session and industrial seminars on vision science.

• 

  Google India

  Aug 2020 - Aug 2020

  Google AI Summer School

  Got selected in highly competitive summer school. Looking forward to connect and learn some excellent brains at Google AI and DeepMind.

• 

  SAP Labs India

  Jul 2021 - now

  • Machine Learning Engineer

    May 2022 - now

  • Associate Software Developer

    Jul 2021 - Apr 2022