Sai Ramachandran

PhD Student at Technical University of Munich

Speaker

Sai Niranjan is a first-principles researcher and PhD candidate at TU Munich, where he investigates the fundamental principles governing the emergence of structure in complex stochastic systems.

His research synthesizes concepts from statistical physics, game theory, and machine learning to find non-obvious, unifying laws in high-dimensional data. His recent NeurIPS paper exemplifies this approach, using fluctuation theory to deconstruct the "black box" of diffusion models and reveal a predictable hierarchy of discrete phase transitions that govern their dynamics.

This drive to bridge foundational theory with practical execution is embodied by his work as a co-founder of a data-focused startup building high-performance infrastructure in Rust. Ultimately, his work is about mastering the full lifecycle of a quantitative idea: from the initial, "alpha-generating" insight to the "operational alpha" of a robust, resilient, and engineerable system.