In a presentation from the "AI Engineer" YouTube channel, Infant and Vaibhav, engineering directors at BlackRock, explain their strategy for **scaling custom AI and knowledge applications** within the world's largest asset management firm. They highlight how **investment operations teams require complex internal tools** for various tasks, including document extraction, workflow automation, Q&A systems, and agentic systems. The speakers outline significant challenges in app development, such as **prompt engineering complexity**, choosing effective **LLM strategies**, and **deployment hurdles** like infrastructure and cost control. To overcome these, BlackRock developed a framework featuring a **"sandbox" for domain experts to rapidly build and refine extraction templates and workflows**, and an **"app factory" for seamless, cloud-native deployment**, drastically reducing development time from months to days. Their key takeaways emphasize the importance of **investing in prompt engineering, educating the firm on LLM strategies, evaluating ROI, and designing with "human-in-the-loop" systems** due to the highly regulated financial environment.