Saurav Sengupta Email & Phone Number

Charlottesville, Va, Us

Charlottesville, Virginia, United States

Currently working at the intersection of healthcare, computer vision and natural language processing. Working with the NIH on characterizing Long COVID (https://ncats.nih.gov/research/research-activities/n3c)

Charlottesville, Virginia Area

• Developed CNN based models to detect gastro-intestinal diseases (Crohn’s, Celiac) on high resolution biopsy images.• Gastroenterology Data Science Lab received $100k Litwin IBD Pioneers Program, Crohn’s & Colitis Foundation grant based on work.

Charlottesville, Virginia Area

Projects:Building CNNs to classify duodenal biopsy images into diseases – Child Health Research Center, UVA• Used Convolutional Neural Networks (CNN) to classify high resolution digitized biopsy images into Celiac Disease, Environmental Enteropathy and Normal tissues. • Used pretrained Resnet50 in our analysis, performs well on diverse set of images.• Achieved close to 98% biopsy level accuracy. Paper accepted in IEEE BHI 2019 conference.• Backed by the Bill and Melinda Gates Foundation and Aga Khan University.• We are also using metabolomics data to find important features of classification.Music Genre Classification• Classification of songs into one of 13 genres like rock/pop and country etc. using ensemble models for song features and lyrical data.• Song data features available from AWS Public Dataset. Used genre data from different Kaggle datasets.• Created AWS EC2 instance, mounted snapshot, ran Jupyter notebooks on EC2 to get data.• Used Random Forests, SVM, Naïve Bayes for model using song features. Used tf-idf and topic modeling for lyrical data. • Ensembled both these models. Around 58.06% accuracy, close to the state of art of 61% but we classify more genres.Modeling Brain Wave Activity using Muse™ headset• Used Logistic Regression for real time classification of brain wave signals like alpha waves into left-right, up-down motion. Tested model by moving objects in a game environment.

Hyderabad Area, India

Working in a project for the Development Bank of Singapore to replace current legacy technology based software with cloud based software using Pivotal Cloud Foundry and 12 Factor based apps using Spring Boot and Spring Data in an AGILE work environment. I was responsible for writing REST API based web services using Spring Boot to be deployed on a Cloud Foundry platform.I was responsible for the following:1. Writing micro-services in Spring Boot to be deployed in Pivotal Cloud Foundry (PCF) platform.2. Setting up various services in the cloud platform like Config Server for parameterizing various application properties and Eureka service registry for central registration of all micro-services.3. Setting up Jenkins CI/CD pipeline for continuous deployment and integration.4. Researched on Elasticsearch for easy cataloging and searching of data.5. Setting up remote Database connections from Pivotal cloud platform to Maria DB server.6. Writing and creating JPA entities.7. Wrote micro-service for sending scheduled emails to parameterized receivers based on the output of a database query using Spring JPA.8. Worked on implementing BPMN2 based workflows on Camunda Modeler for business process automation of bank reporting processes.

Greater Noida

I interned on the team responsible for designing a embedded neural network ASIC (Application Specific Integrated Circuit).My responsibilities included:• Writing RTL (Register Transfer Level) design using Verilog HDL of APB (Advanced Peripheral Bus) slave. It involved going through the documentation from ARM to understand how to implement the architecture.• Inserting reset synchronizers in the design.• Interfacing random access memory with different peripherals.• Using Synopsys Design Compiler to synthesize RTL code to netlist. Since we write the implementation of the ASIC using Hardware Description Languages such as Verilog and VHDL, we need to convert such code into gate level designs for actual implementation of the logic in silicon, which is called as synthesis. This is then verified using test cases for various inputs.• Inserted Design for Testability (DFT) in the design using Synopsys DFT compiler. Also inserted externally controllable resets in the design.• Inserted On Chip Clock controller using DFT Compiler.• Formally Verified design using Synopsys Formality.I wrote Tcl/Tk and Unix shell scripts for Synopsys Design Compiler, DFT Compiler and Formality for accomplishing the above tasks.

Interacted with company officials on how they are implementing WCOM (World Class Operations Management)in the Tetra Pak, Chakan plant.Learnt the 5S methodology and its implementation in an industrial environment.

Designed a working system to monitor and report the temperature of oil flame in the thermal power plant.Learnt various processes taking place inside a thermal power plant, with focus on electronic instrumentation inside the power plant.

Learnt about various technologies applied in Control and Navigation of aircraft focusing on Doppler VHF Omni-directional Radio Range, InstrumentLanding System and RADAR.Learnt about various communication systems used during Air Traffic Control.

Doctor Of Philosophy - Phd, Data Science, Machine Learning For Healthcare, Computer Vision, Nlp

Master'S Degree, Data Science

Activities and Societies: Best Poster Award - 2019 IEEE Systems and Information Engineering Design Symposium

Bachelor Of Engineering (B.E.), Electrical And Electronics Engineering

Activities and Societies: Nirmaan Goa