Sushant Singh

• Leadership and Strategic Influence: Orchestrated design and development of complex forecasting and optimization models, impacting multi-million dollar business decisions. (e.g., Force To Load - Network Management Center project)
• Advanced Analytics Expertise: Developed and deployed machine learning models using techniques like Prophet, Boosting, NLP, and n-gram analysis for demand forecasting, fraud detection, customer experience scoring, and churn reduction. (e.g., Vendor Fraud Identification, QES (Quality Experience Score), Digital Upgrade Propensity projects)
• Full Project Lifecycle Management: Owned projects from inception to deployment, including data engineering, model building, visualization, maintenance, and business communication.
• Measurable Business Impact: Delivered solutions that resulted in significant cost savings, revenue growth, and improved customer experience. (e.g., Force To Load project saved $10M, QES project increased customer acquisition by 10-20%)
• Technical Breadth and Depth: Proficient in various programming languages & frameworks (Python, PySpark, Kedro), cloud platforms (GCP, Domino, Yahoo), data visualization tools (Looker, Tableau), and NLP techniques.
• Mentorship and Knowledge Sharing: Built reusable tools and APIs for data exploration and model performance analysis, benefiting multiple teams. (e.g., Auto EDA Forecasting Tool project)

• Predictive Maintenance: Built a highly accurate (90-95%) WTTE-RNN model to predict bleed fault failures in aircraft 10+ flight cycles in advance, saving significant costs and downtime (Bleed Fault Failure prediction in Aircraft).
• Travel Propensity Modeling: Proposed and developed a multi-class, multi-label classification model predicting customer travel propensity for a major Telecom client. This unlocked new revenue streams by identifying potential high-value customers (Customer Travel propensity for next destination).

Strong Business Acumen and Delivery:

• Increased Revenue by $2.4M/Year: Developed innovative API for a Telecom client to predict customer travel propensity, creating new revenue opportunities by aligning business KPIs (Customer Travel propensity for next destination).
• Improved Operational Efficiency: Built a linear ensemble ML model for automotive client that accurately predicted vehicle ETA (R-square: 0.98, RMSE: 2.8), significantly optimizing logistics planning (Vehicle ETA prediction).

Leadership and Problem-Solving:

• Oversaw Anomaly/Fraud Detection: Orchestrated development of a VAE-ANN model achieving F1-score of nearly 0.98 for a highly imbalanced dataset, effectively identifying fraudulent warranty claims (Anomaly/Fraud Detection).
• Project Ownership and Transition: Successfully transitioned a critical platinum application for a major Automotive client in Europe, providing ongoing support and enhancements (Automotive - Vehicle logistics & Invoicing).

Technical Breadth and Versatility:

• Experience with Diverse Techniques: Possesses expertise in various methods including RNNs, ensemble learning, anomaly detection, LSTMs with attention, CNNs, and NLP techniques.
• Experience with Big Data Platforms: Successfully deployed a model on Skywise (Airbus) big data platform using PySpark (Bleed Fault Failure prediction in Aircraft).