Resume – Varun Singh

Resume – Varun Singh#

[alt text]

A data-driven professional with a strong foundation in Data Science and Civil Engineering, currently pursuing a Master’s degree in Data Science at the University at Buffalo. Skilled in programming languages such as Python, SQL, Java, and JavaScript, with expertise in data management, analytics, and visualization tools including PowerBI, Tableau, and advanced statistical techniques.

Experienced in developing predictive models and machine learning solutions across diverse projects, with a focus on the pharmaceutical industry. Demonstrated proficiency in forecasting, automation, and data-driven decision-making through roles at Viscadia and IQVIA. Adept at leveraging statistical techniques and machine learning algorithms to extract meaningful insights from complex datasets, to enhance business processes and strategic decision-making.

Currently expanding knowledge in the field of data science while applying analytical and problem-solving skills to real-world challenges. Committed to driving innovation through the strategic application of data science methodologies, particularly in the pharmaceutical and healthcare sectors. Passionate about exploring the intersection of data science and healthcare to enhance patient outcomes and optimize healthcare delivery systems.

Education#

Master of Science: Data Science, University at Buffalo, The State University of New York: Fall 2024 - Fall 2025

Skills & Tools#

  • Languages: Python (Numpy, Pandas, Scikit-learn, Keras), SQL (MySQL, PostgreSQL, MongoDB), Java, Javascript, R

  • Data Management & Analytics: Apache Spark, Apache Kafka, Amazon Sagemaker, PowerBI, Tableau

  • Web Development: React.js, Node.js, Flask, FastAPI, RESTful API, HTML, CSS

  • Tools: Jupyter, Visual Studio, MATLAB, MS Azure, Render, Docker, Git

Work Experience#

Associate - Forecasting, Viscadia, Gurugram, India: Dec 2023 - Jul 2024#

  • Built forecasting models using time series and scenario-based techniques for pharmaceuticals in the EU and Canada

  • Standardized 30+ models to enhance accuracy and address regional market variations

  • Automated VBA modules for sensitivity analysis and Monte Carlo simulations, reducing project timelines by 50%

  • Leveraged APLD, claims data, and industry sources (Biomedtracker, EvaluatePharma) to deliver actionable industry insights

Data Analyst, IQVIA, India: Aug 2022 - Nov 2023#

  • Automated reporting with Python (Matplotlib, Pandas, NumPy, Scikit-learn), saving 10+ hours weekly and improving accuracy

  • Developed a web tool for RFP project, automating data processing and integrating authentication

  • Led a PowerBI initiative tracking brand performance, optimizing data from 10+ sources, adding 40+ billable hours, and earning client recognition

  • Managed Above-Country engagement for 20+ European countries, ensuring timely, accurate report delivery with minimal errors

  • Created a PBI dashboard and Python backend, automating analysis of promotional data from 10+ sources

  • Co-authored a respiratory market case study, covering SDR, forecasting, HCP segmentation, and ROI analysis for launch decisions

  • Mentored team members on reporting, ad-hoc analysis, and IQVIA data platforms

Projects and Technical Papers#

Heart Disease Prediction Model:#

  • Developed an end-to-end ML pipeline for classification, from data preprocessing to model deployment

  • Implemented feature engineering, selection, and PCA for performance optimization

  • Experimented with multiple classifiers, logging results via MLflow on DagsHub

  • Deployed the model as a containerized FastAPI application with a Streamlit interface

  • Documented the process comprehensively in a Jupyter Book with video presentation

Restaurant Website:#

  • Built a full-stack restaurant website using Node.js, Express.js, and PostgreSQL with JWT authentication.

  • Implemented responsive frontend design with dynamic menu display and reservation system.

  • Developed RESTful API endpoints for user authentication, menu management, and reservation handling.

  • Designed PostgreSQL database schema with relational tables for users, menu items, and reservations.

  • Deployed and managed application across Render (frontend/backend) and Koyeb (database) platforms.

Crime Detection using Sentiment Analysis#

  • Developed a hybrid crime prediction model integrating sentiment analysis from social media with historical crime data.

  • Implemented data collection pipeline using Playwright and PRAW, processing over 10,000 tweets and 20GB of Reddit data

  • Utilized NLTK and TextBlob for sentiment analysis, achieving 87% accuracy in classifying social media posts

  • Created custom scoring mechanism assigning severity scores (1-10) to different crime types based on their impact

  • Integrated sentiment analysis results with historical data modeling to generate cumulative safety scores for states

  • Validated model results against established crime indexes from Wikipedia, USA News, and Numbeo

Generative AI for Mental Health Support:#

  • Designed a generative AI-powered web application for therapeutic suggestions and mindfulness exercises using GPT-4 and Hugging Face Transformers.

  • Developed a responsive frontend using React.js with features for emotion-based therapeutic responses and user-customizable tone preferences.

  • Built a Flask/FastAPI backend to process user inputs and integrate AI-generated outputs dynamically.

  • Implemented privacy-focused safeguards with no data storage and automatic redirection to professional resources when needed.

  • Leveraged MongoDB for storing non-sensitive configuration data and deployed the application on AWS and Render platforms.