* Identified high-impact opportunities with XFN partner org to leverage LLM-powered automation and assistance within their human review labeling workloads. Delivered $4M+ annual OpEx savings.
* Spearheaded the conception and delivery of an AI copilot agent for human review workloads, leveraging LLM models to enhance human annotation quality via fact-checking and guideline-based improvements, image and video content understanding with vision models, queries against a knowledge base with RAG, policy clarification etc. Built a self-serve integration platform, saving $6M+ OpEx in 0.5 years.
* Initiated a capacity efficiency initiative to reduce training wait times amidst capacity constraints.
