At Google [X], Yash and his team faced the intricate challenge of extracting actionable insights from human brain data—a process inherently difficult due to its availability, noise and complexity. To tackle this, they devised innovative methods, including synthetically generating brain data and feature engineering algorithms that localized time-series signals from specific brain regions into independent AI models (such as auditory and visual). They fine-tuned these models for accurate inference scores and even identified attention shifts between different brain regions.

After transitioning to Facebook Libra and exploring blockchain technologies like Bitcoin, Yash recognized similar challenges in global financial markets. These markets were also plagued by noisy time-series data, limited historical records (about ~50 years), and spatial characteristics that hindered the performance of AI models in predicting flow of funds (liquidity) across different asset classes.

In 2022, inspired from his experiences at Google [X] and Facebook Libra, Yash began building AI models for financial assets as a weekend project. By late-2023, the results were so compelling that he teamed up with longtime friend Amit to launch Highwater AI.