Sarah Al-Said sarahal-said

About

sarah = {
    "role"       : "Software Engineer",
    "focus"      : ["Prompt Engineering", "RAG Architecture", "LLM Evals", "AI Safety"],
    "stack"      : ["Python", "LangChain", "LangGraph", "FastAPI", "AWS"],
    "experience" : "software engineering · healthcare operations",
    "education"  : "BS + MS Computer Science",
    "location"   : "South Florida · Remote",
    "status"     : "Open to Software Engineering roles",
}

I build machine learning systems that are reliable by design — eval-first development, guardrails from day one, observability as a first-class concern. With software engineering experience across financial services, healthcare, and marketing technology, and a background in regulated healthcare operations, I build AI that works in the real world.

What I Build

🔍 RAG Pipelines Hybrid retrieval with BM25 + dense search, reranking, citation tracing, and hallucination guardrails. Not toy demos — production systems with measurable accuracy improvements.	🧪 LLM Evaluation Eval frameworks that treat prompts like code. Offline regression tests, online A/B experiments, CI/CD gates that block deploys on faithfulness regression.
🛡️ AI Safety Infrastructure Drop-in LLM proxies with PII redaction, toxicity filtering, red-team test suites, and real-time observability dashboards for production LLM I/O.	⚡ Prompt Systems at Scale Prompt architecture that's versioned, measured, and production-hardened. System prompts, few-shot templates, dynamic context injection — all treated as engineered artifacts.