I am currently working as a Full-Stack Software Developer at Aeroleads. However my expertise and experience goes well beyond development.
I love getting my hands dirty in a variety of things, such as Robotics, Computer Vision and Deep Learning, Game Development, Building Simulations, and 3D Modelling. I am a firm believer in building user-centric products and designing before coding.
For more information of each award, click on their images
Awarded by Akilur Rahman, CTO of ABB for winning Smart India Hackathon Software Edition 2018. Project Page
Other Awards:
Runner’s up, ML Sprint BVP
Winner HackBVP 2.0
Runner’s up, Build For India -Paytm Ideathon Awarded by the Vice President of Paytm
Winner, NIEC Hackathon
Clicking on individual projects will lead you to their own page
(* work in progress)
Egocentric to bird's eye view for vehicles *
Proposed an approach to transfer the egocentric views from a vehicle (for example from a dashcam) into the bird's eye view of the environment around the vehicle.
Advised by Dr. Vanita Jain
Simultaneous Pose and Appearance transfer
Devised a novel approach to achieve pose and appearance transfer simultaneously using only a single image of the person. Our model is trained in an adversarial manner using a cyclic consistency loss.
Advised by Dr. Koteswar Rao Jerripothula
3D reconstruction from egocentric cameras
Devised a novel approach to reconstruct the mesh and the textures of the user's body from cameras installed under the VR headset. Our method makes use of the available methods for reconstruction by generating an intermediate hollistic view from the egocentric views.
Advised by Dr. Vanita Jain
Tensorflow Blog | Poster | Video | Marconi Society Blog | Business Wire
Built a deep learning driven android application which can estimate the air quality (PM2.5 value) around you using a single image. Uses Federated Learning to improve the model’s performance while preserving the privacy of the users as the user base grows.
Awarded by Mrs. Padmasree Warrior (ex-CTO, Cisco and Motorola) for winning Paul Baran’s Young Scholar’s Celestini Prize India, 2019 @ IIT Delhi. Our work was also featured on the Tensorflow Medium Blog.
Advised by Dr. Aakanksha Chowdhery (Google Brain, Tensorflow) and Dr. Brejesh Lal (IIT Delhi)
An autonomous line following bot that uses a path following algorithm to traverse. Using a camera at the top of the room, the bot and other artifacts on the arena weredetected and animated models were displayed on them using augmented reality. Link to interview.
First Prize in e-Yantra 2018 @ IIT Bombay
A humanoid robot to help nurses do muscle intensive tasks like lifting and turning patients by mimicking their actions. The nurse wears a vest embedded with motion sensors and the motion is refined and further sent to the humanoid
First Prize in SIH 2018 Hardware Edition
A machine learning model that outputs an optimal price for charging stations based on several parameters such as the time and the day, the number of cars in the station, traffic around the station, etc. Uses the YOLO object detector for detecting cars through survellience cameras. I also built an android application for manually adjusting the parameters and enabling personalization for the model.
First Prize in SIH 2018 Software Edition
Shivam Grover, Kshitij Sidana, Vanita Jain, ”Egocentric Performance Capture: A Review”, Fusion: Practice and applications (FPA), Volume 2 , Issue 2, PP: 64-73 , 2020.
Vanita Jain, Shivam Grover, Kshitij Sidana, G.T Thampi, Achin Jain, and Manju Khari, ”Pipeline for 3D reconstruction of the human body from AR/VR headset mounted egocentric cameras”, IEEE Transactions on Computational Social Systems, vol. 7, no. 4, AUGUST 2020 (Submitted)
Shivam Grover, Koteswar Rao, ”Controlled Simultaneous Pose and Appearance Manipulation”, (Unpublished)
Research Intern at NICE Labs @ IIIT Delhi (Present)
Working on a novel approach for performing realistic human pose and appearance in images and videos. Advised by Dr. Koteswar Rao Jerripothula
Research Intern at IIT Delhi (May 2018- October 2018)
Worked on a deep learning driven android application which can predict the air quality (PM2.5 levels) around you using a single image of your environment. Advised by Dr. Aakanksha Chowdhery (Google Brain, Tensorflow).
Our poster won the second prize and was awarded by Mrs. Padmasree Warrior.
