Sharjeel Imtiaz
Lead Data Scientist Architect @Wipro
Signup · Get unlimited contacts
WORK HISTORY
Lead Data Scientist Architect @Wipro
Client : State Street Bank Led a cross-functional team of 4–5 to design and deliver a scalable legal document key term extraction and Q&A system using Databricks, Azure Gateway, and a Bronze–Silver–Gold RAG-based architecture- Built and deployed a Retrieval-Augmented Generation (RAG) pipeline for legal document understanding with context-aware prompt engineering and factual consistency checks, enabling high-precision automated legal insights- Developed a custom evaluation framework (RAGAS alternative) incorporating prompt-based contextual metrics — Contextual Precision, Faithfulness, and Coherence — and implemented a ground truth labeling process to ensure consistent benchmarking- Managed project delivery, evaluation cycles, and stakeholder reporting, defining key tasks, metrics, and quality gates across data engineering, NLP modeling, and evaluation streams- Proposed Advance routing based architecture which routes multiple sources using query, LLM conjecture pipeline- Designed a reusable chatbot framework adaptable across multiple enterprise domains (Legal, Finance, HR), promoting code reuse and rapid deployment- LangChain tools for Entity or meta building, LLAMA index for chunking strategies, LangChain pedantic for intent, metadata, robust planner match with query to answer- Reduced manual legal document review time by ~70%, improving operational efficiency and compliance accuracy- Improved Q&A contextual accuracy to 85%+, validated against labeled ground truth datasets- Increased model evaluation speed and reliability by 60% through automation of contextual metrics and dashboards- Established a scalable and reusable AI evaluation and chatbot framework, now serving as a baseline for future enterprise AI initiatives- Expert of pydantic and LangChain Agentic workflow for extraction, RAG, and multi-source environments and tools
EDUCATION
University of East London
Doctor of Philosophy (Ph.D in NLP), Data Science
Arid and Agriculture University
MCS, Computer Science/Database
International Islamic University, Islamabad
MS, Database /Data Analysis
ABOUT SHARJEEL IMTIAZ
With over 17 years of professional experience—more than a decade of which has been dedicated to AI and Machine Learning—I have evolved from foundational data roles to strategic leadership as a Data Scientist and Machine Learning Engineer. I design, build, and deploy cutting-edge AI models and end-to-end ML pipelines that deliver measurable business outcomes.ML and DomainsBuilt classification, regression, clustering, and deep learning models using XGBoost, SVM, ANN, Auto-Encoders, Transformer-based models, and Reinforcement Learning.Designed and deployed low-latency streaming inference systems for document Q/A using Kafka + feature store for train/serve parity, with PSI-based drift detection and auto-retraining triggers. Leveraged PyTorch/TorchScript for optimized, high-throughput inference.I lead initiatives in Anomaly Detection, Root Cause Analysis (RCA), Supply Chain Forecasting, A/B Testing, Marketing Mix Modeling Sales Forecasting, NLP-based RAG services, Demand Forecasting, Elasticity Modeling, and Aspect Mining for Tourism Recommendation Systems.Applied Causal inferring using A/B testing to uplift the profit over organic search vs paid one for Podcast.Developed predictive models in Python (scikit-learn, XGBoost) to estimate customer lifetime value (CLTV) using CTR, sCPM and identify high-propensity paid users; improved targeting precision and increased conversion rates by 10% for games day.I bring deep expertise in developing scalable machine learning solutions and AI applications, with domain proficiency in Retail Insights, LangChain, and Large Language Models (LLMs). My work spans the full ML lifecycle, including cloud deployment, AI robustness testing, and model operationalization, all while fostering strong cross-functional collaborations.Led coaching like role for A/B testing across projects to validate recommendation systems, chatbot enhancements, and modeling.Working experience on Agents workflows like LangChain with multiple agents design pattern and reflection pattern experience, automated RAG intent, query enhancement and retrieval scoring using Agentic workflow with external tool of email.Fine-tuning experience of LORA, Q-LORA based method and Databricks experience with Genie where deployed model for extraction.Extensive experience with NLU/NLP data, including LLMs, sarcasm detection, and document chunking.Developed prototype conversational AI experiences (Q/A bots, semantic search, personalized recommendations) in collaboration with design, ad-tech, and product teams.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.