Jan 2022 – Apr 2022 · 3 mos · Jodhpur, Rajasthan, India · Remote

• On Behalf of NVIDIA, conducted a Hands-on Lab course on Deep Learning Operations (DLOPs): CSP 7030 for a capacity of 130+ students from various streams of B.Tech and M.Tech.
• Curriculum:-
• 1. Overview of Computing Infrastructure
• 2. Basics of Docker and Containers
• 3. Deep Dive into Kubernetes
• 4. Overview of AI and Deep Learning
• 5. Hands on with Deep Learning Training Operations on Multi GPU/ Multi Node environment
• 6. Hands on Deep Learning Optimization
• 7. Hands on Deep Learning Inference Deployment

Jun 2019 – Mar 2020 · 9 mos · Raipur, Chhattisgarh, India · On-site

• Biomedical Imaging and Deep Learning-based Project at "Electronics and Communication Engineering, NIT Raipur"
• Mentor Prof.:  Dr. Shrish Verma
• Deep Dilated V-Net for 3D Volume Segmentation of Pancreas in CT images (Accepted at SPIN-2020, IEEE International Conference)
• Proposed Deep Dilation Network (DDN) extracts spatial contextual information by capturing different receptive fields.
• To avoid class-imbalance, we proposed a Weighted Fusion Loss (WFL). Results are – Dice Score: 83.3%, Sensitivity: 87.7%, & Precision: 83.0%.
• Language: Python3, Framework: PyTorch, Tools: ITK-Snap, MITK

May 2019 – Jul 2019 · 2 mos · Bengaluru, Karnataka, India

• Computer Vision Project at "Video Analytics Lab, CDS, Bangalore"
• Mentor Prof.: Dr. R. Venkatesh Babu
• Earned a fellowship amount of 25,000 INR by Indian Academy of Sciences, Bangalore ( SRFP, IAS Fellowship )
• 1) Retrieving Representative Samples from a trained classifier (Under Submission)
• Using kernels of a prior pre-trained model, implemented a data-free GAN for generating representative samples.
• Verified the samples by achieving an accuracy of 71.12% by student network on the Knowledge Distillation task.
• Language: Python3, Framework: PyTorch

Jan 2019 – Aug 2020 · 1 yr 7 mos · Singapore, Singapore · Remote

• Deep Learning-Based Projects at "Medical Mechatronics Lab, NUS".
• Mentor PhD:   Dr. Mobarakol Islam
• Mentor Prof.:  Dr. Ren Hongliang
• 1) RNA Coding Sequence Estimation Using BSTA Module (Under Submission)
• Implemented a novel embedding method & a Bottleneck Spatio-Temporal Attention (BSTA) module incorporated with Resnet18.
• We achieved a SOTA accuracy 88.9% & specificity & 89.0%, for the classification of code & non-code sequences.
• Language: Python3, Framework: PyTorch

Sep 2018 – Sep 2019 · 1 yr · India

• Our device works in two modes. The first mode is scene and object detection in which it detects the objects in front of it, calculates the distance of the object from it and gives the person an auditory response. By this, the visually impaired are able to detect the objects and obstacles. The second mode is that in which the device recognises the printed text where the finger points and converts it into audio so that the blind person is able to hear the text. In this way, our device helps the visually impaired to act independently by working in two modes. There is no such innovation prior in the technology world serving these purposes.We are developing a prototype of the above mentioned device.

May 2018 – Jul 2018 · 2 mos · Tirupati Area, India

• Image Processing Project at "Electrical Engineering, IIT Tirupati"
• Mentor Prof.: Dr. Rama Krishna Gorthi
• Eigen Vectors based Rotation Invariant Multi-Object
• Deep Detector
• ( Paper Accepted at SPIN-2020, IEEE International Conference)
• Designed a novel Eigen Vectors based pre-processing and a decision criterion step for a deep neural network object detector to detect rotated objects in images.
• For 3 differently oriented angles, 90°, 180°, and 270°, and shows a high accuracy boost by 48%, 50%, and 47% respectively, over normal YOLOv3 ( Without explicitly training with augmented dataset)
• Language1: Python3, Framework: PyTorch, Language2: MATLAB

Dec 2017 – Jan 2018 · 1 mo · Ghaziabad, Uttar Pradesh, India

• Briefly studied about different Telecommunication topologies, & different generations of mobile communication Along with Next generation Network, Broadband, Wifi, Wi-Max, LAN, WAN, OFC etc., and Basics of IP Addressing (IPv4 and IPv6) at BSNL- ALTTC.