About me

Hi there! I'm Suvrat Jain, a year old data scientist with a strong background in machine learning, data analytics, and software development. I'm a problem solver at heart and my passion for data science stems from the desire to use technology to tackle real-world challenges.

I hold a Master's degree (M.Sc.) in Data Science from Rochester Institute of Technology, NY and I'm always looking for new and exciting ways to grow my skills and contribute to the field.

I fell in love with the world of code at the young age of 17 and my fascination with technology has only grown stronger since then. With a creative mind and a love for finding patterns in data, data science was the natural next step for me.

I've worked on a variety of projects that showcase my skills in machine learning, computer vision, data analytics, and software development, and I'm always eager to take on new challenges and push the boundaries of what's possible. Some of these projects can be found here.

My 5 year goal is, to successfully marry...my love for building delightful websites and the knack for solving problems using data, to birth a data-driven product that helps in the greater good of the community.

I am constantly looking to enhance my skill set and work on new & exciting projects, and have a track record of being a reliable asset to both - myself & the organization I work with.

🙋🏻‍♂️If you are a company/organization looking to hire an enthusiastic individual for an exciting job opportunity, to contribute to your rapid growth, feel free to reach out to me via LinkedIn.
👋🏻Or if you want to drop a hi or brainstorm something that interests you about music, universe, data, or anything in between, feel free to reach out to me here. Check out my resume!

education

M.Sc. (Data Science)

Rochester Institute of Technology, NY, United States
2021 - 2023

Courses: Applied Statistics, Java & Python for Data Science, Neural Networks, Time Series Analysis & Forecasting, Visual Analytics

B.Tech (Electronics & Communication Engineering)

Manipal Academy of Higher Education, Dubai, U.A.E.
2017 - 2019

Courses: Embedded Systems, Analog Communication, Digital Signal Processing using MATLAB, VLSI Network Design

work experience

Data Science Engineer

Golisano Institute for Sustainability, RIT

Rochester, New York, United States
May 2023 - Present

  • Led the development of a vision system integrating YOLOv8 and UNet segmentation models to achieve precise clothing dismantling and optimize downstream processes.
  • Leveraged AWS SageMaker to build scalable image classification pipelines using TensorFlow. Implemented distributed training across multiple instances for improved efficiency.
  • Implemented advanced data augmentation techniques, resulting in a 10% improvement in IoU score for the semantic segmentation model, showcasing expertise in enhancing data quality and accuracy.
  • Developed and built ETL and data augmentation pipelines, to expand dataset by 30% and achieve a 12% lift in model accuracy due to improved generalization.
  • Implemented efficient CUDA kernels for GPU acceleration, demonstrating proficiency in deep learning frameworks and performance optimization.
  • Deployed a vision system prototype, configuring industrial grade cameras and lighting and establishing TCP/IP network communication for seamless integration with a control system.

    • Data Scientist

    Golisano Institute for Sustainability, RIT

    Rochester, New York, United States
    May 2022 - May 2023

  • Developed performant CNN models for automobile part categorization (98.2% on 10k+ images, using OpenCV). Demonstrated expertise in techniques with broad AI applications, including image processing and transfer learning.
  • Boosted automobile part classification accuracy by 73% by fine-tuning an Inception v3 model (TensorFlow), demonstrating skill in transfer learning for rapid solution development.
  • Implemented a Siamese Network for similarity learning, combining distance metrics and statistical analysis techniques. Achieved 98% accuracy on out-of-training data, significantly improving model generalization.
  • Developed a multi-stage ETL pipeline for object detection/classification for a sustainable remanufacturing vision system, reducing manual inspection time by 30% and boosting object classification accuracy by 10%.
  • Successfully translated complex data insights into clear and impactful visualizations, effectively communicating key findings to a diverse audience of stakeholders, including C-level executives, software developers, and product managers.
  • Enhanced ML pipeline efficiency through targeted research and experimentation (A/B testing, Bayesian optimization). Built clear data visualizations (Tableau, Power BI) to convey model behavior and drive improvements, presented to senior leadership. (Project Demo)

    • Software Developer and Tester

    Merkle

    Pune, Maharashtra, India
    September 2020 - August 2021

  • Deployed 100+ email campaigns for brands like Estée Lauder and MAC Cosmetics using Salesforce Marketing Cloud, delivering an average click-through rate of 25%.
  • Developed and delivered personalized email campaigns with a reach of more than 400,000 customers in 100+ countries worldwide, using Braze.
  • Developed automated test suites (Python, Selenium) for web applications, reducing manual testing effort by 24% and accelerating release cycles.
  • Built API test automation scripts to validate functionality and data integrity. Increased test coverage by 25% for core back-end services.
  • Performed SQL queries to extract and analyze user behavior data from Adobe Analytics and Salesforce Marketing Cloud.
  • Leveraged Adobe Analytics to identify key audience segments and target them with relevant ads, for 50+ email marketing campaigns.

    • QA Associate

    Merkle

    Pune, Maharashtra, India
    January 2020 - September 2020

  • Reduced overall test cycle time by 40% using Python scripting in automation testing.
  • Identified & reported 500+ unique bugs across web & mobile applications for UltraTech Cement, in what was to be one of the most successful launches of the company's digital initiative.
  • Performed manual testing of mobile applications, including smoke, functional, integration, regression, user acceptance, and UI testing for multiple releases each sprint.

    • Web Developer (Intern)

    Arachnomesh Technologies Pvt. Ltd.

    Navi Mumbai, Maharashtra, India
    July 2018 - September 2018

  • Accelerated page load times by ~22% through optimization of CSS, HTML, and JS.
  • Designed web application components in Angular JS, including login interface and dashboard views.
  • Developed 10+ responsive web pages for clients in a variety of industries, utilizing HTML5, CSS3, JavaScript and jQuery to design interactive user experience.

    • Software Developer (Intern)

    G10 Consultancy Digital Pvt. Ltd

    Mumbai, Maharashtra, India
    May 2017 - July 2017

  • Implemented 10+ web designs in a mobile-first, responsive approach for optimal viewing across browsers and devices.
  • Worked on WordPress CMS, HTML, CSS, JavaScript, and Bootstrap framework to build websites and apps for 10+ clients.

    • SKILLS

    Python

    Java

    C++

    TensorFlow

    PyTorch

    Open CV

    Pandas (Python)

    NumPy (Python)

    SQL

    AWS

    Jupyter

    Minitab

    Salesforce

    HTML5

    CSS3

    JavaScript

    Angular JS

    WordPress

    PROJECTS

    Predicting Hospital Readmissions: A Machine Learning Approach with Logistic Regression, Decision Tree Classifier, and Random Forest Classifier

    Binary classification using machine learning models to predict if hospital patients are readmitted within 30 days. Models: Logistic Regression, Decision Tree Classifier, Random Forest Classifier.

    Eventify-U

    Discover your next event, your way! Personalized recommendations, hassle-free searching, and endless possibilities with Eventify-U!

    Winning the Game with Data: An Insightful Exploration of IPL 2008-2020

    Exploratory data analysis carried out on the IPL dataset for seasons 2008 - 2020. Data visualizations used to extract insights and factors contributing to victories and defeats.

    Moodytector

    Moodytector is a simple web application that detects facial features and emotions. Face-api JS was used to build the majority of this app. There are three main features: emotion detection, age prediction, and gender prediction. The site dynamically updates on user input through webcam.

    Get Insight into User Engagement with HUE Vision: Real-Time Eye-Tracking Web Application

    A web application built using TensorFlow.js to detect and track eye movement through webcam. Application: Analyze users' gaze movement on your website in real time using a webcam. It can be useful for UX researchers when doing usability studies to understand how users respond to the aesthetics of your website.

    EmotionWise: Unlocking the Language of Faces with Real-Time Emotion Recognition

    EmotionWise is a web application built with Affdex SDK by Affectiva that does real-time emotion recognition to learn the language of faces by visualizing its emotions in real time. Key features include: expression, appearance, glasses, gender & age.

    Unlock Your Account with a Smile: A Facial Authentication Web App

    Facial authentication web app that detects faces and compares them with your previously stored facial data, if a match is found it logs you into your account. Built using face-api.js.
    Try out the demo below!

    Building Bridges: Sri Sri University's Model United Nations Website

    The Model United Nations is one of Sri Sri University's annual events which provides a platform for students to engage in meaningful discussions to solve real world problems and also learn about underrepresented nations and cultures. This is a responsive website created to showcase and provide information about the event. Built using HTML, CSS, JS.

    Spread Smiles, Generate Joy: A Compliment Generator Built with Love

    Is there something which is keeping you upset or low? Worry not, am here to stay & cheer you up for as long as you wish.
    This is a simple JS based compliment generator that gives you genuine compliments to cheer you up and hopefully, to make your day.
    Made with ♥ using JS, HTML, CSS.

    Pico.js Face Detection: A Simple JavaScript Demo Project

    This demo project shows how to use pico.js to create a simple application for face detection using JavaScript.