Sanchit Thakur

I’m a 2nd-year CS (AI/ML) student specializing in probabilistic ML, deep learning, and autonomous systems. I focus on distribution-aware thinking and hard validation, not just model accuracy, so the system stays reliable when the world changes.

As the independent researcher behind the Empirical Survivability Framework (ESF), I designed a fully distribution-free probabilistic system for non-stationary sequential environments. ESF validated across 18/18 regime shifts in equity, cryptocurrency, and climate domains with zero parameter retuning, and I proved key results like an optimal observation window and explicit trust convergence bounds.

I also build production-grade ML systems and applied deep learning to real-world workflows. In “Chaching,” an autonomous multi-broker trading bot, I integrated FinBERT for real-time multimodal signal fusion and built an async data pipeline with caching, retry logic, and a live Flask dashboard tracking PnL and risk metrics.

Beyond research and engineering, I lead and collaborate—currently as the Head – Pixel Club (AI/ML) with 50+ members and as a founding core member of the ACM Student Chapter. I’m eager to keep growing where probabilistic ML, deep learning, and autonomous decision-making can be shipped with strong experimental discipline.