Targeted LLM Workflows Reduce Hallucination Risk

Apurv Jain is the Founder and CEO of MacroX, a San Francisco-based startup building a real-time measurement layer for the global economy using alternative data. Before MacroX, Apurv built machine-learning investing systems at Microsoft, traded FX derivatives at Deutsche Bank, and worked at Bridgewater Associates. His research experience spans Microsoft Research and Harvard Business School, with publications and invited talks across academia, central banks, and global forums including Davos and the World AI Summit. He holds degrees from Georgia Tech and the University of Chicago and is passionate about mixed martial arts. In this podcast, we discuss:

1. Official Data Lags & Declining Survey Response Rates
2. 1,000+ Alternative Data Pipelines for Macro State Mapping
3. Real-Time Worker Bargaining Power vs. Lagged Wage Metrics
4. Multimodal AI: CNNs, BERT, and Targeted LLM Workflows
5. Nowcasting Lead: Gaining 2–3 Months Ahead of Official Prints
6. Next-Generation Nowcasting vs. Atlanta Fed GDP Models
7. Why the 2023 Yield Curve Recession Signal Failed
8. Current US Economic State & The "K-Shaped" Consumer
9. Portfolio Integration for Fundamental & Systematic Investors