About Me
I graduated as an interdisciplinary MS Scholar in electrical and computer science engineering department at IIT-Madras and my guide at IIT-M was Dr. Kaushik Mitra. Previously, I had completed my B.Tech Degree with honors in Mechanical Engineering in 2017 from NIT-Raipur.
I am interested to work in the areas of AI/ML where it can have a positive impact in the society. So after graduation, I joined as a data scientist in a healthcare product based company named dozee in Bangalore to work in their remote patient monitoring product and currently working as a data scientist at Tredence Analytics, Bangalore. Apart from my interests to explore the related areas of AI/ML, I am also interested to explore the areas related to Quantum Computing and AI.
Publications
High-Speed HDR Video Reconstruction from Hybrid Intensity Frames and Events. [Paper] [Slides]
Accepted in the International Conference on Computer Vision and Machine Intelligence (CVMI), 2022
Rishabh Samra, Dr.Kaushik Mitra, Prasan Shedligeri.
Clinical Validation of an Indigeneous Micro-Vibration Vital Parameter Monitor Dozee VS [Paper]
Accepted in the International Conference on COMmunication Systems & NETworkS (COMSNETS), 2023
Accepted in the International Conference on COMmunication Systems & NETworkS (COMSNETS), 2023
WORK EXPERIENCE
DATA SCIENTIST , TREDENCE ANALYTICS, BANGALORE
Following are the major projects I am involved in:
Migration of 30+ ML pipelines from kubeflow to databricks for a US-Based firm.
A GenAI-based solution of document summarization for a US-based Private Equity Firm. This solution resulted in 80% reduction in document generation time, has potential to optimize team size from 20 to 5 members resulting in estimated annual cost savings of $4.5 Million.
Causal Discovery from the structured IID observational data.
Applying machine learning predictive models to patient history data for hospital admission/ readmission prediction, ER Risk prediction and hospice condition probability estimation.
Given the EMR lab test data, we build RoBERTa based classification models for filtering incorrect lab tests. Then hierarchical clustering algorithms was used to cluster similar labs names that can be integrated for enhanced analysis.
Uncertainty Quantification in Neural Networks.
DATA SCIENTIST, DOZEE, BANGALORE
The work involves building AI-driven statistical and parametric models to derive meaningful features and information from the complex signal (ballistocardiography and electrocardiography signals). Following were the major projects I was involved in:
Developed and tested an efficient algorithm to calculate the Respiratory Rate vitals from the Ballistocardiography signals. Testing of data was done on 8000 hrs. of data in uncontrolled settings and 584 hours of data in controlled settings.
Analyzing sleep patterns from ballistocardiography(BCG) data for automated sleep staging.
Created a pipeline for the analysis of over 14000 Heart Rate (HR) and Respiratory Rate (RR) vital points for getting US-FDA 510K clearance.
Creating a model for signal quality classification from the ECG data that will be helpful for Arrhythmia Detection pipeline. [Ongoing]
Apnea/Non-Apnea classification from the BCG signal using the methods of deep learning [Ongoing]
Projects
I illustrated Curriculum Learning on custom made lines' and characters' datasets where the motivation comes from learning happening in children who first learns to draw simple lines and progressively learns about characters. [Report]
High Speed HDR Image Frame Reconstruction from Event Sensor Data . [Preliminary Slides]
Given patient level data with their characteristics and another data containing time series biometric information of the patients, we aim to reduce patient mortality by identifying high risk patients based on the available data and devoting the limited ventilators to them. [Details]
Experimental Investigation of Stress Concentration Factor in a rectangular plate with central hole, using Photoelastic Stress Analysis and simulating results using ANSYS. [Report]
Designing an efficient algorithm to calculate the respiratory rate from the Ballistocardiography signals received via Dozee device.
Presentations
Articles and Research Paper Summaries
Understanding Information-Seeking, Curiosity and Attention [Write-up]
Causal Learning in Children by Cristine Legare [Paper Summary]
Jain Theory of Mind [Article & Paper Summary]
Recent Achievement and Workshop Attended
Active contributions to the AI-CoE team at Tredence Analytics resulted in receiving the Founder’s Award. [Certificate]
Received Pat On the Back award from Tredence for contributing in one of the highly influential GenAI project of Document Summarization. [Certificate]
Attended workshops related to the Future of Quantum Information Processing organized by IIIT-Dharwad. [Certificate]
Attended workshop sessions on Quantum Computing organized by Association of All Computer Science Teachers. [Certificate]