SHUBHAM PRATAP SINGH
Data Scientist/ML Engineer
About
AI/ML Engineer and Data Scientist with 2+ years of experience delivering production-grade machine learning, forecasting, anomaly detection, and Generative AI solutions across supply chain and manufacturing domains. Experienced in building scalable data pipelines using Python, PySpark, Azure, and Databricks, developing predictive models, RAG systems, and LLM-powered applications. Proven track record of delivering business impact through forecasting, optimization, and AI-driven automation initiatives.
Skills & Expertise (26)
Work Experience
Consultant – AI/ML/Data Science
Ernst & Young LLP
Mar 2024 - Present
Developed a scalable ingestion pipeline for a Multimodal RAG application, integrating Azure Blob Storage, APIs, and direct user uploads. Implemented AI-powered metadata enrichment and document summarization using Azure OpenAI GPT-5.1 on Microsoft Foundry. Built a context-aware extraction and semantic chunking framework using LangChain and Docling to preserve document structure and meaning. Enabled enterprise-scale semantic search through embedding generation with text-embedding-3-large and indexing with Azure AI Search. Developed and implemented an agentic chat layer for the RAG application, enabling LLM-driven query understanding, intelligent routing, tool invocation, and grounded response generation using LangGraph. Implemented dynamic retrieval and tool selection based on query intent, supporting reasoning over text, tables, images and PDF content. Designed multi-step agent workflows and conversational context management to handle complex and multi-turn user queries. Evaluated and optimized agentic RAG responses using relevance, faithfulness, and completeness metrics.
Consultant – AI/ML/Data Science
Ernst & Young LLP
Dec 2025 - Apr 2026
Built a scalable anomaly detection and data quality framework for large-scale (BigData) supply chain dataset up to 450 million rows, using statistical techniques including Z-Score and IQR analysis, improving data reliability for downstream analytics. Applied time-series clustering to identify operational patterns and segment similar business behaviors, enabling more accurate anomaly identification and trend analysis. Implemented advanced imputation strategies for missing and corrupted records, achieving 80% reconstruction accuracy.
Consultant – AI/ML/Data Science
Ernst & Young LLP
Mar 2025 - Nov 2025
Extracted, cleansed, and transformed workforce attendance data from enterprise SQL databases, ensuring high-quality datasets for downstream analytics and model development. Conducted exploratory data analysis across multiple manufacturing plants to identify attendance trends, seasonal patterns, and operational drivers, leveraging insights for feature engineering and feature selection. Developed and benchmarked multiple regression models, improving manpower demand prediction accuracy enabling production workforce planning across multiple manufacturing plants. Designed and deployed a production-grade workforce forecasting solution achieving 95% prediction accuracy.
Consultant – AI/ML/Data Science
Ernst & Young LLP
Sep 2024 - Mar 2025
Developed an end-to-end lead and transit time prediction solution using large-scale supply chain transaction data from vendors and manufacturing plants, including data extraction, cleansing, and exploratory analysis. Analyzed shipment and transaction completion patterns to identify key business drivers influencing delivery performance, enabling targeted feature engineering and selection. Implemented statistical outlier detection and mitigation techniques using Empirical Rule analysis to improve data quality and model reliability. Evaluated and benchmarked multiple regression models to identify the optimal forecasting approach.
Consultant – AI/ML/Data Science
Ernst & Young LLP
Mar 2024 - Sep 2024
Built a multi-country demand forecasting solution for the Asia-Pacific region by integrating sales, market, economic, geographical, and pandemic-related factors into predictive models. Applied multiple advanced time-series to improve forecast accuracy for monthly and annual planning. Evaluated and selected the best-performing forecasting models through comprehensive benchmarking and performance analysis.
Education
Post Graduate diploma in Artificial Intelligence - Center for Development of Advance Computing
2023 - 2024 · Afghanistan
Master of Engineering/Technology, Computer Science & Engineering (AI/ML) - Rajiv Gandhi Proudyogiki Vishwavidyalaya
2019 - 2021 · Afghanistan
Bachelor of Engineering, Computer Science & Engineering - Rajiv Gandhi Proudyogiki Vishwavidyalaya
2012 - 2016 · Afghanistan
Google Certified Cloud Digital Engineer - Google Cloud Platform
- 2026 · Afghanistan
Certifications
Google Certified Cloud Digital Engineer
Google Cloud Platform · 2026
Interested in this developer?
Profile Score Breakdown
Profile Overview
Availability Details
Visa Status
Citizen
Relocation
Depends on Offer
Skills (26)
Click a skill to find developers with the same skill
