Haseeb Zahid

• Implemented Retrieval-Augmented Generation (RAG) with vector databases, enabling 40% more accurate and contextually relevant LLM responses. • Developed a multi-functional AI chatbot using LangGraph with support for tools, in-memory and SQL-based persistent memory, Human-in-the-Loop (HITL), time-travel, and custom session states, enabling contextual, accurate, and history-aware conversations. • Implemented scalable workflows and memory management, integrating API/tool chaining, session tracking, and histor

• Engineered a transformer-based SQL translation model (PostgreSQL → MSSQL) using TensorFlow, achieving >90% translation accuracy. • Designed and automated end-to-end AutoML pipelines with Scikit-learn, TPOT, and PyCaret, improving fraud detection accuracy by 18%. • Built and deployed scalable Flask, Django, and FastAPI web apps for real-time ML inference. • Streamlined CI/CD pipelines with Docker, Jenkins, and GitHub Actions, cutting model deployment times by 40%. • Developed transformer-based text summari

• Built and managed PySpark and SQL ETL pipelines for genomic datasets exceeding 10TB, enabling faster genome-wide association studies (GWAS). • Designed and implemented ML workflows using Random Forest, Gradient Boosting, and Neural Networks to identify patterns in health datasets. • Applied BERT and ALBERT NLP models to extract and classify medical text with 92% precision. • Deployed containerized ML models on Docker and orchestrated workflows with Jenkins, ensuring reproducibility and scalability.