Stephen Varghese

About Stephen Varghese

Data Science Researcher focused on solving complex problems at the intersection of machine learning, physics, and engineering. Extensive experience at Shell translating advanced research into tangible impact, including the development of AI-powered systems for anomaly detection & predictive maintenance, ML models for time series forecasting & production optimization, and statistical data fusion frameworks for GHG accounting & data reconciliation.Expertise lies in bridging the gap between advanced algorithms and real-world operational efficiency to make industrial processes safer, cleaner, and more productive. Rigorous academic foundation from IIT Delhi (B.Tech), UIUC (MS), and TU/e (PhD).

Experience

1. Data Science Researcher

   Shell

   Nov 2018 — Present · Bengaluru, IN

   Development and deployment of advanced data science solutions using machine learning and statistical modelling to solve complex challenges in the energy sector. Focused on sensor fusion, anomaly detection, predictive maintenance, time series forecasting, dynamic process modelling, supply chain and process optimization to deliver operational improvements across Shell assets.Key Achievements:• Sensor Fusion & Soft Sensors: Architected a data fusion framework for reliable mass balance and GHG accounting. Co-developed an ML-based soft sensor to quantify flare emissions for regulatory audits. Deployed in two LNG assets.• Anomaly & Fault Detection: Deployed a real-time leak detection system for cryogenic heat exchangers, now critical for turnaround planning in four LNG assets. Developed novel algorithms for control valve stiction detection.• Predictive Maintenance: Built data reconciliation frameworks to detect sensor bias, enabling predictive maintenance for production critical sensors. Deployed at multiple Shell assets.• Product Optimization: Contributed towards development of ML models to maximize LNG production efficiency by optimizing control variable set points. Deployed in one LNG asset.• Supply Chain Optimization: Contributed to a supply chain optimization framework using MILP for optimal distribution of low-carbon fuel facilities.• Advanced Root Cause Analysis R&D: Statistical modelling and causal information theory to isolate root causes of equipment failure. Spearheading R&D using Graph Neural Networks (GNNs) for plant-wide anomaly prediction.• Advanced Dynamic Modelling R&D: R&D on Physics-Guided AI to build deep learning-based dynamic process models for transient process conditions in heat exchangers.

Education

• Indian Institute of Technology, Delhi

  Bachelor of Technology (B.Tech.), Mechanical Engineering

  2006 — 2010

• University of Illinois at Urbana-Champaign

  Master of Science (MS), Mechanical Engineering

  2011 — 2013

• INSA Lyon - Institut National des Sciences Appliquées de Lyon

  Bachelor of Technology (B.Tech.), Mechanical Engineering

  2008 — 2009

• Eindhoven University of Technology

  Doctor of Philosophy (PhD), Applied Physics

  2014 — 2018