Eman Fatima Artificial Intelligence Engineer & Cybersecurity Analyst

Architected and deployed a tamper-proof decentralized hospital management ecosystem (healerplus.live). Engineered Ethereum smart contracts (Solidity) for cryptographic Role-Based Access Control (RBAC) and immutable audit logging, integrated IPFS decentralized storage for patient electronic medical records (EMR) with zero single-point-of-failure, and embedded an AI Random Forest anomaly detection engine that actively evaluates user access metrics and flags malicious intrusion attempts in real time.

Web platform that analyzes uploaded audio to detect synthetic/tampered deepfakes, using speaker diarization, temporal inconsistencies, timestamps, and metadata analysis to generate full forensic audit reports.

Full stack platform detecting phishing across URLs, emails and web content in real time using NLP, WHOIS and DNS lookups, and outputting confidence scored risk classifications.

Deep learning model trained on the German Traffic Sign dataset to detect and classify road signs using computer vision techniques with 98%+ accuracy.

Records user recitation audio and evaluates phoneme-level pronunciation against renowned Quranic reciters, pinpointing exact mispronounced words with accuracy feedback.

Architected an end-to-end data analytics and predictive benchmarking pipeline. Formulated advanced data preprocessing workflows handling skewed distributions, missing value imputation, multi-collinearity checks (VIF), and interquartile outlier filtering. Benchmarked 6 machine learning algorithms (Random Forest, XGBoost, SVM, Gradient Boosting, KNN, and DBSCAN for spatial density clustering) with rigorous 10-fold cross-validation, hyperparameter tuning via GridSearchCV, and interactive Seaborn/Plotly visual diagnostics.

Designed a deep learning computer vision system for real-time multiclass animal classification. Engineered a custom multi-layer Convolutional Neural Network (CNN) with Batch Normalization, Dropout regularization, and Softmax probability distributions. Implemented an innovative Confidence Threshold Gate that flags out-of-distribution (OOD) and ambiguous inputs to prevent false positive classifications, wrapped in an interactive Streamlit UI with real-time class heatmaps and confidence score telemetry.

Engineered a high-performance credential security auditing platform enforcing modern NIST SP 800-63B compliance standards. Implemented mathematical entropy analysis, dictionary attack pattern detection, and real-time credential verification against compromised database records using k-Anonymity SHA-1 cryptographic range hashing—ensuring zero sensitive plaintext exposure. Features comprehensive password health scoring, policy enforcement, and forensic compliance auditing.