Hire Destynee - Prompt Engineer - San Francisco

Architect and deploy production-grade prompts and advanced AI agents within the real estate division. Oversee multi-agent execution pathways and prompt version control to actively mitigate operational risk and prevent model hallucinations or bias within financial and underwriting data.

Contributed to the design and development of structured, reasoning-based intelligence systems. My work includes prompt engineering and creating evaluation methodologies and benchmarks to measure reasoning consistency, fairness, and interpretability in Large Language Models. Our lab investigates critical model failure modes including hallucination and weak faithfulness to improve the reliability and transparency of modern LLM/AI architectures.

Improved model alignment and reliability by evaluating reasoning quality, auditing supervised fine‑tuning datasets, and conducting comparative scoring to strengthen reinforcement learning from human feedback.