Curriculum Vitae
Resume pdf
Education
- Ph.D in Computer Science, Indiana University Bloomington, 2021 - 2026 (expected)
- GPA: 4/4
- Courses: Algorithm Design and Analysis, Programming Language Principles, Explainable AI, Computer Networks
- M.S. Data Science, Indiana University Bloomington, 2019 - 2021
- GPA: 3.97/4
- Courses: Deep Learning, Advanced Natural Language Processing, Machine Learning, Elements of Artificial Intelligence, Introduction to Statistics, Advanced Database Concepts, Exploratory Data Analysis, Computer Vision
- M.S in Data Science (Data Analytics Track) with thesis
- Master’s thesis: Response-Based Knowledge Distillation
- B.Tech in Electrical Engineering, National Institute of Technology (NIT) Patna, India, 2011 - 2015
Research Experience
- Research Assistant, Indiana University Computer Vision lab
- Deep Learning (DL) - Case-Based Reasoning (CBR) survey paper: June 2020 – Present
- Collaborated with the DL-CBR research group in writing two sections of the survey paper- retrieval and DL models with memory.
- Exploring Accuracy and Explainability of DL-CBR Hybrid Architectures: Jan 2022 – Present
- Exploring hybrid systems leveraging knowledge-engineered and network-learned features in concert.
- Analyzing different DL architectures to study the impact of learned features in the DL-CBR hybrid system.
- Multi-View Stereo (MVS) Method for High-resolution Depth-map Prediction and Point Cloud Generation: Aug 2021 – Present
- Improving the application of the MVS algorithm for high-resolution depth map prediction
- Applying the MVS algorithm at multiple stages of DL architecture to improve the accuracy and completeness of 3D point cloud
- Roof-area Segmentation and Orientation Detection: Aug 2021 – Present
- Detection of roof orientation and plane area on 3D point cloud data by applying the RANSAC algorithm
- Improving roof segmentation and orientation detection using satellite images
- Master’s Thesis on “Response-based Knowledge Distillation”: Aug 2020 – May 2021
- Analyze the knowledge distillation process and propose the soft-label hypothesis to explain the behavior of the distillation process.
- Proposed special consideration for pre-training teacher models for retaining similarity information in soft labels for better knowledge distillation.
- A condensed version of the thesis was published at the AAAI-2022 workshop
- External Research Fellow, NIT Patna (Electrical Engineering Lab) : July 2018 – June 2019
- Project title: “Sustainable Smart Grid Framework for Energy Management System Incorporating Available Renewable Resources.” funded by the Science & Engineering Research Board, Government of India.
- Successfully conceptualized and implemented a model to mitigate the Communication-link failure in a smart meter-based load forecasting system using various classification methods.
- Implemented an electrical load forecasting system with one year of data using a polynomial regression model.
- Presented paper at the 4th IEEE International Conference on Computing Communication and Automation (2018).
Publications
Work Experience
- Course Development and Instructor – Computer Vision Paper Discussion Section, Indiana University: Spring 22 and 23
- Designed and was the instructor of the paper discussion section, the first of its kind at IU, in the computer vision course
- Lead the discussion of DL-based seminal papers covering the development of CNNs, Transformers and MLP-based models
- Received Associate Instructor of the Year award, 2021-2022 for my distinguished contribution to this course
- Associate Instructor, Indiana University
- Computer Vision: Spring 2021
- Elements of Artificial Intelligence: Fall 2020, Spring 2021
- Held weekly office hours on Zoom to work one-on-one with undergraduate, and graduate students and working professionals.
- Manage and help more than 100 graduate students with their doubts about Piazza/InScribe.
- Senior Manager, Tata Motors Ltd. Pantnagar. 2015 – 2017
- Analyzed and optimized the maintenance frequency of Generator yard equipment using past breakdown and maintenance data.
- Overhauled and systematized power transformer oil filtration frequency of 36 units of transformer, individually, by predicting oil characteristics threshold value against number of operation hours of each transformer using eight years of past records.
Teaching
Skills
- Languages & Tools: Python, SQL, R, PyCharm, Rstudio, PostgreSQL, C
- Frameworks & Libraries: Pytorch, TensorFlow, Keras, NumPy, Pandas, NLTK, Pillow, Scikit-learn.
- Certifications:
- “Machine Learning” and “Deep Learning” specialization by Prof. Andrew Ng on Coursera,
- “TensorFlow in practice” by Laurence Moroney from Google.
- “Statistical Learning” by Prof. Hastie and Prof. Tibshirani on Stanford online
Leadership and Awards
- AI of the Year award, Luddy School of Informatics, Computing, and Engineering, 2021-2022 for exceptional contribution to teaching.
- Best Graduate (Boy) National Institute of Technology Patna, batch of 2015.
- Honored by Tata Sustainability group for distinctive work in CSR under Pro-engage program in 2016-17.
