Bengaluru, Karnataka, India
I’m a Data Scientist with experience building production-grade ML systems across supply chain, retail/CPG, and e-commerce. Over the last few years, I’ve worked end-to-end across every layer of the analytics lifecycle, right from problem framing and raw data engineering to model development, deployment, and business consumption. My focus areas include: - Demand forecasting & supply planning - Predictive modelling & decision systems - Inventory & replenishment optimisation - MLOps, automation & data pipelines - Business analytics & stakeholder enablement I enjoy solving messy real-world problems where accuracy, impact, and long-term reliability matter. I’m fluent in Python, SQL, PySpark, scikit-learn, TensorFlow/Keras, MLflow, Airflow, and Databricks and I care as much about clean, maintainable code as I do about model performance. At my core, I love using data to build systems that make businesses faster, smarter, and more resilient.
Specialized in predictive modeling for end-to-end AI-powered SaaS solutions supporting global CPG and retail clients, enabling smarter, faster, and more profitable decisions across the supply chain lifecycle. Key Contributions & Impact Demand Forecasting Platform (Production Ready & Scalable) → Led development of a forecasting engine ingesting sales, product attributes, promotions, holidays/events, and external factors. → Clustered thousands of SKUs by forecastability (forecastable, semi, non-forecastable), ensuring tailored modeling strategies. → Built and deployed multiple model families: → ANN + LSTM for complex/intermittent patterns → LightGBM with enriched engineered features → Triple Exponential Smoothing (TES) for stable series → Delivered a blended ensemble approach with rolling window validation → Result: ~20% improvement in forecast accuracy vs naive baselines → Integrated into enterprise planning tools across global CPG accounts MLOps & ML Engineering Improvements → Developed modular, reusable Python and PySpark pipelines on Databricks. → Introduced MLflow tracking and Model Registry, improving transparency and collaboration & Implemented Hydra-based configuration management, reducing iteration cycles. → Automated workflows using Airflow, enabling reliable retraining, monitoring, and alerting →Designed a PySpark + pytest validation framework, improving deployment stability and code quality
Troubleshooting complex Content Delivery Network issues, related to Application layer protocols, Origin Storage, Streaming and other OTT technologies.