profile picture

Shashank Shekhar

Masters student in AI

Education

MASc in Engineering (Aritficial Intelligence specialisation)

University of Guelph, Canada
GPA : 90.14/100
2019 - present

BTech in Electronics And Communication Engineering

Indian Institute of Technology, Dhanbad
GPA : 8.17/10
2013 - 2017

Honors & Awards

NeurIPS Travel Grant

2019

ICCV Student Volunteer Award

2019

ICCV Travel Grant

2019

JN Tata Travel Grant

2019

JN Tata Endowment for Higher Education of Indians

2019

Vector Institute Scholarship in AI

2019

Machine Learning Summer School London Full Scholarship

2019

Indian Academy Of Sciences Summer Research Fellowship

2015

Coursework

Graduate

Introduction to Machine Learning
Computational Methods for Aritficial Intelligence
Optimization
Scientific Computing
Computational Statistical Inference
Machine Learning for Natural Language Processing

Summer Schools

Machine Learning Summer School 2019 London
Deep Learning For Bayesian Methods Summer School 2019 Moscow

Online

Machine Learning
Deep Learning Specialization(5 course specialization)
Practical Reinforcement Learning (with honors)
Mathematics For Machine Learning Specialization (3 course specialization)

Technical Skills

Languages

Python, C, C++, Matlab, Julia

Machine Learning Frameworks

PyTorch, TensorFlow, Scitkit-Learn, NLTK

Computer Vision Frameworks

OpenCV, Torchvision

Development Tools

Scientific Python Stack, Jupyter, Sublime Text, Vim

DeveOps Tools

Docker, AWS, Git, JIRA

Academic Research Experience

Graduate Research Assistant

January 2020 - Present
Machine Learning Research Group, University of Guelph and Vector Institute of Artificial Intelligence

Pursuing research on explaining and interpreting deep visual representations using methods from cognitive sceince

Project Research Assistant

January 2018 - April 2019
Visual Computing Lab, Department Of Computational And Data Sciences, Indian Institute Of Science

Worked on different computer vision research and engineering problems both in academic and industrial projects. Topics included object detection, person re-identification and visual question answering

Summer Research Intern

May, 2015 - July 2015
Indian Institute Of Technology, Hyderabad

Implemented the following algorithms to simulate an Internet-Of-Things HetNet in MATLAB: Down-link proportional-fair and SCFDMA Up-link proportional-fair scheduling for LTE, Page-segmentation and RAW based channel-access scheduling for 802.11ah.

Industrial Research Experience

Research Associate

October 2018 - April 2019
Shell Technology Center, Bangalore

Implemented image denoising, contrast enhancement, segmentation and super-resolution algorithms for large scale 3D µ-CT digital rock images (MATLAB, C++)

Computer Vision Intern

March 2018 - September 2018
Hyperworks Imaging, Bangalore

Implemented person detection, labeling and re-identification methods on real-world videos from a multi camera network (Pytorch, OpenCV, PyQT5)

Publications

From Strings to Things: Knowledge-enabled VQA Model that can Read and Reason
Ajeet Kumar Singh, Anand Mishra, Shashank Shekhar, Anirban Chakraborty
IEEE International Conference on Computer Vision 2019 (Oral: 4.3% acceptance rate)

OCR-VQA: Visual question answering by reading text in images
Anand Mishra, Shashank Shekhar, Ajeet Kumar Singh, Anirban Chakraborty
International Conference on Document Analysis and Recognition 2019

Operator-In-The-Loop Deep Sequential Multi-camera Feature Fusion for Person Re-identification
Navaneet K L, Ravi Kiran Sarvadevabhatla, Shashank Shekhar, R. Venkatesh Babu, Anirban Chakraborty
IEEE Transactions on Information Forensics and Security (Volume: 15)

Road Damage Detection And Classification In Smartphone Captured Images Using Mask R-CNN
Shashank Shekhar and Janpreet Singh (equal contribution)
IEEE International Conference On Big Data Cup 2018 Challenges Workshop

Software Development Experience

Content Engineer

April 2019 - June 2019
Deeplearning.ai

Developed Tensorflow assignments and Docker based auto-graders for Coursera’s Deep Learning specialization

Software Development Engineer

July 2017 - January 2018
SVC-SW Team At Samsung Research Delhi

Worked on the video-on-demand applications team to develop streaming applications like Netflix and Amazon Prime for Linux-based Tizen-OS for Smasung's Smart TV platform

Data Science Intern

June, 2016 - September, 2016
BlackCoffer

Worked on individual projects consulting on text analysis and time series modelling

Projects

Deep Policy Gradients To Train Musculo-Skeletal Controller To Walk - Reinforcement Learning based controller for musculo-skeletal model to enable it to walk using policy gradient approaches (Proximal Policy Optimization (PPO) and Deep Deterministic Policy Gradients (DDPG)). Received $500 cloud credits from Google for being Top 100 in NIPS 2018 challenge
Playing FPS Game from Raw Pixels by Combining Improvements From Deep Q-Learning - Combined improvements in Deep Q-Learning like target nets, dueling architecture, prioritized experience replay etc as in state-of-the-art RAINBOW (paper from DeepMind at AAAI 2018) to play the FPS game DOOM using the VizDoom framework from only the image pixels.
Bi-Linear Convolutional Neural Nets For Fine-Grained Visual Classification - Implemented Bi-Linear CNNs for Fine-grained Visual Recognition (ICCV 2015) and ensembled with other classifiers to perform both single-label and multi-label type image classification forthe iMaterialist Challenge as part of the FGVC5 Workshop at CVPR 2018. Finished <50/~1000 participants
Classifying Hate Speech in Wikipedia Dialogues using Recurrent Neural Nets - Implemented Long Short Term Memory (LSTM), Gated Recurrent Units (GRU), Bilinear & Stacked versions of both, Attention as well as using word embeddings from Glove and FastText to perform text classification. Achieved 98.58% mean ROC-AUC score

Teaching

Graduate Teaching Assistant

January 2020 - April 2020
Engineer Analysis ENGG*1500, University of Guelph

Graduate teachign assistant for undergraduate course with over 200 students on linear algebra and its applications in engineering. Designed and taught the module on ordinary least squares method.

Community Teaching Assistant

June 2018 - Present
Machine Learning, Coursera

Serving as a community TA on the Coursera Machine Learning after being invited for exception performance and participation in the course and completing Coursera's two week training on the same. Involved with solving learner's issues related to the topics and programming assignments via discussion forums as well as mediating with Coursera for the same.

Extracurricular Activites

Member

September 2017 - December 2017
Patent Ideation Team, Sasmung Research Institute - Delhi

Worked to develop novel ideas for application of AI techniques to multimedia streaming for the SVC-SW team.

Member and Co-Ordinator

August 2014 - May 2017
Quiziapa (Quiz Club), IIT (ISM) Dhanbad

  • Organised college level quiz events and participated in inter-college and national level events