Nada Bastawi

I've built and shipped AI products including AgentForge TriageCrew, a CrewAI-based multi-agent platform that automated GitHub issue triage end to end and achieved 19/19 task completion on a live FastAPI benchmark.

For TORCO Pest Control, I took a hybrid recommendation engine from concept to production, combining content-based and SVD collaborative filtering. I solved cold-start with customer and property features, reached Precision@5 above 95%, and deployed natural-language access through FastAPI and an MCP server.

I've also designed an Arabic RAG system using Pinecone and Chroma, and architected a multimodal ICU outcome-prediction system combining clinical text, OCR, and physiological signals. My capstone work included preprocessing MIMIC-III signals, generating mortality labels from ICD-9 codes, and presenting model trade-offs to technical and clinical faculty.

In contract and trainee engagements, I've delivered industrial object detection and computer-vision attendance systems from requirements through client delivery. I work across Python, ML/DL frameworks, FastAPI, CI/CD, monitoring, containerization, and cloud deployment, and hold the Oracle Cloud Infrastructure Generative AI Professional certification.