نبذة عني
AI Engineer with hands-on experience building enterprise AI solutions using Python, LLMs, RAG, AI agents, OCR, and NLP. Experienced in developing FastAPI AI microservices, bilingual Arabic/English document-intelligence w…
AI Engineer with hands-on experience building enterprise AI solutions using Python, LLMs, RAG, AI agents, OCR, and NLP. Experienced in developing FastAPI AI microservices, bilingual Arabic/English document-intelligence workflows, structured extraction, vector search, and human-in-the-loop AI systems. Strong foundation in Machine Learning, Deep Learning, and Computer Vision, with practical experience integrating AI services with REST APIs and event-driven backend systems.
الخبرة
NLP Intern / AI Projects
<p>Built and evaluated RNN/LSTM text-classification models; the selected LSTM achieved ~97% validation accuracy, 0.9398 Macro F1, and 0.9619 Weighted F1.</p><p> Developed a Streamlit demo combining LSTM inference with BLIP image captioning for text and image classification; explored INT8 quantization and memory efficiency.</p>
AI Engineer Trainee
<p>Develop and support AI-powered solutions for regulatory compliance, policy management, document intelligence, and privacy/data-protection workflows.</p><p> Build AI-agent workflows using Gemini, prompt engineering, structured outputs, confidence scoring, grounding, validation, and human-review guardrails.</p><p> Implemented semantic obligation matching using sentence-transformer embeddings and Qdrant vector search for policy-management workflows.</p><p> Developed bilingual Arabic/English PDF processing with PyMuPDF and EasyOCR fallback for extracting obligations, deadlines, recurrence, citations, and compliance metadata.</p><p> Integrated Python/FastAPI AI microservices with backend services through REST APIs and Kafka-based asynchronous workflows; used Git, Bitbucket, Jira, Postman, Docker, and pytest.</p>
المشاريع
RRM AI Service – Multi-Agent Regulatory Intelligence
<p>Developed an enterprise AI service for regulatory relationship management using a multi-agent architecture. The system includes specialized agents for document ingestion, AI-assisted reply drafting, obligation extraction, and matter summarization.</p><p>Built bilingual Arabic/English document-processing workflows with hybrid PDF extraction and OCR fallback for scanned documents. Implemented structured outputs, confidence thresholds, validation guardrails, and human-in-the-loop review.</p><p>Technologies: Python, FastAPI, Google Gemini, PyMuPDF, EasyOCR, REST APIs, Pydantic, Kafka, Docker, and AI agent workflows.</p>