Safwa Jabbar

I’m an Applied AI Engineer with 3+ years of experience designing and deploying machine learning and generative AI systems. I focus on end-to-end, agentic workflows—LLM orchestration, retrieval augmented generation (RAG), and multi-agent pipelines—so outputs become actionable decisions, not just prototypes.

At U: The Mind Company, I designed and shipped an N8N agentic workflow for Patent Prior Art search and evaluation, using structured LLM outputs and automated scoring logic. I also built prompt-engineered GPT-4 workflows that interpret EEG signals and auto-generate structured medical diagnostic reports, achieving 85% concordance with neurologist assessments. Alongside that, I developed an end-to-end EEG ML pipeline with custom preprocessing (Gaussian filtering, CAR), improving artifact removal efficiency by 30%.

I’ve strengthened clinical-grade signal and perception pipelines using computer vision and workflow automation. Using MediaPipe and YOLO for Parkinson’s disease detection, I improved hand landmark detection accuracy by 25%, and I collaborated cross-functionally to validate AI outputs for 95% clinical relevance. I also built CRM-integrated agentic workflows connecting Salesforce and HubSpot via REST APIs to automate lead enrichment, activity logging, and structured data extraction for GTM operations.

Beyond production systems, I develop robust agentic research and tooling—like an AI-Powered Academic Peer Review Pipeline (N8N + Claude) with GROBID XML parsing, 7 parallel reviewer agents, structured JSON outputs, and Supabase score aggregation. In parallel, I research privacy-preserving, HIPAA-compliant methods for sensitive medical video data and identified movement biomarkers with 92% accuracy, while continuing to push reliable orchestration using guardrails and error handling. My certifications and recognition (Salesforce Platform Developer I, hackathon finalist/top performer) reflect how I combine engineering rigor with practical delivery.