LinkedIn | pranavsp108@gmail.com | GitHub | Kaggle

I am a Master of Science in Analytics candidate at the University of Minnesota - Twin Cities, graduating in May 2026. Most recently, as a Data Science and Optimization Consultant at Daikin Applied, I engineered a predictive framework that identified over $1.5M in annual manufacturing cost savings and reduced simulation runtimes by 90%.

My professional background includes over 2.5 years at Tata Consultancy Services as a Consultant on data and optimization projects, where I automated mission-critical reporting for global retail operations and orchestrated large-scale cloud migrations. I specialize in building end-to-end solutions, from automated ETL pipelines to complex predictive models using Python, SQL, and MLOps frameworks.

I am currently focused on full-time opportunities in Data Science and Machine Learning Engineering, specifically within Predictive Modeling, Recommendation Systems, and Operations Research.

• Predictive Modeling & Optimization: Scikit-learn, XGBoost, LightGBM, TensorFlow, Causal Inference, Recommendation Systems, Time-Series (LSTM), Statistical Modeling, A/B Testing.
• Optimization & Operations Research: Linear Programming (PuLP, SciPy), Simulation Modeling, Constraint Optimization, Supply Chain Analytics.
• Data Engineering & Analytics: Advanced SQL, Python (Polars/Pandas), PySpark, Databricks, dbt, Airflow, Kafka, ETL/ELT Pipeline Development.
• MLOps & Cloud: MLflow, Docker, Kubernetes, GitHub Actions (CI/CD), AWS (S3, EC2, SageMaker), Azure, GCP (BigQuery, Vertex AI), Model Monitoring & Governance.

GitHub Repo

• Problem: Financial market data is highly volatile, making static forecasting models obsolete within days.
• Solution: Engineered an automated pipeline using Kafka and GitHub Actions to ingest 26+ years of data into an S3 Data Lake. Developed a stacked LSTM Neural Network on AWS EC2 with a "zero-touch" MLOps workflow for daily retraining.
• Impact: Fully automated the end-to-end forecasting of 9 global market indices with real-time data ingestion and model monitoring.

[GitHub Repo]

• Problem: Digital platforms often struggle with "information overload," leading to low user engagement and retention.
• Solution: Engineered a robust PySpark ETL pipeline and trained a Scikit-learn collaborative filtering model. Containerized the model using Docker and deployed it as a REST API on Azure for real-time inference.
• Impact: Successfully demonstrated a 25% increase in user engagement metrics through personalized content delivery.

[GitHub Repo]

• Problem: Inefficient inventory management leading to high carrying costs and order fulfillment delays.
• Solution: Developed advanced time-series forecasting models to predict demand patterns across diverse product categories.
• Impact: Projected a 15% reduction in inventory costs and improved order fulfillment rates by 28%.

Beyond these highlights, I have developed 5+ additional projects covering Sentiment Analysis, Statistical A/B Testing, and Supply Chain Simulation.

Programming & Core Tools

Machine Learning & Statistics

Big Data & MLOps

Databases & Cloud

Data Visualization & BI

Data Visualization & BI