Khatija Begum

• Designed and deployed a scalable Healthcare Knowledge Assistant using Azure-based LLMs (OpenAI GPT-4, Bison), improving document search accuracy by 25%.
• Developed and trained ML models (Random Forest, XGBoost, LSTM) on FHIR patient data for risk prediction and time series forecasting, achieving 20% improvement in prediction accuracy.
• Applied NLP (spaCy, NLTK, TextBlob) on medical documents for sensitive entity extraction and sentiment classification, enhancing patient privacy compliance.
• Built and automated ML pipelines using Azure ML, Flask APIs, Terraform, Jenkins, and GitLab CI/CD, maintaining 99% uptime in production.
• Conducted feature engineering and PCA to optimize model performance on large-scale healthcare datasets.
• Implemented Retrieval-Augmented Generation (RAG) techniques using Elasticsearch and Cognitive Search to boost response relevance in healthcare Q&A.
• Led integration of RESTful APIs and IAM roles to support secure, real-time data access in HIPAA-compliant environments.
• Documented technical workflows and maintained project tracking using Git, Bitbucket, Confluence, and Jira under Agile/SCRUM

Environment: Python, Pandas, NumPy, Scikit-learn, TensorFlow, Keras, Azure ML Studio, Azure Functions, Azure Blob Storage, Terraform, Flask, Jenkins, GitLab, Jira, spaCy, NLTK, TextBlob, Power BI, REST APIs

• Developed automated credit risk and churn prediction models using Python (Random Forest, Logistic Regression), improving model accuracy by 18%.
• Created financial dashboards with Tableau and SQL, enabling 30% faster reporting for risk teams.
• Led ETL development using Informatica PowerCenter (SCD Types 1 & 2, CDC) for reliable data warehousing and history tracking.
• Built and optimized data models, performed data validation, and coordinated project rollouts using Agile and cross-functional collaboration.
• Conducted hypothesis testing (A/B, ANOVA) and power analysis to validate customer behavior trends.
• Enhanced data reporting accuracy through automated validation pipelines integrated with Tableau and database queries.

Environment: Python, SQL, Informatica PowerCenter, Tableau, Oracle, AWS Redshift, Agile/SCRUM

• Built and deployed ML models for student churn prediction and behavior segmentation using Scikit-learn, improving student retention strategies.
• Designed ETL pipelines with Azure Data Factory and Databricks to process large-scale user interaction logs.
• Developed NLP models using NLTK and spaCy for sentiment classification and text analytics on student feedback.
• Engineered features using PCA, scaling, and label encoding to prepare clean, high-quality training datasets.
• Delivered actionable insights through interactive dashboards (Tableau, ggplot2) and advanced statistical analysis.

Environment: Python, Pandas, Scikit-learn, Azure Data Factory, Databricks, NLTK, spaCy, Tableau, SQL, R

• Predicted patient readmission risks by training ML models (Logistic Regression, SVM, Gradient Boosting) on EMR datasets with >12M records.
• Built data preprocessing pipelines with Azure Data Factory and performed advanced EDA using Pandas, Matplotlib, and Seaborn.
• Deployed ML models using Azure ML Studio and Azure Functions for real-time inference, reducing manual triage effort by 30%.
• Applied PCA and feature scaling techniques to reduce dimensionality and improve model training performance.
• Integrated model APIs into hospital systems and maintained secure data access through Azure SQL and IAM best practices.

Environment: Azure ML Studio, Azure Data Factory, Azure Functions, Python (Pandas, Scikit-learn, Matplotlib), Tableau, Azure SQL, EMR data