Arnav Dhiman

About

Who Am I?

I am drawn to the edges between fields. The place where mathematics starts to sound like biology, where a proof technique from graph theory says something true about probability, where the way a neural network fails tells you something fundamental about optimization.

Research

Research Intern at MIT CSAIL, Kellis Lab. Nine outputs across AI, biology, quantum computing, and mathematics. Emergent Ventures Grantee. Patent filed.

Background

Based in Mumbai. BeyondQuantum Research Scholar (60/434). Engineering Intern at SkyFlux. IRIS NextGen Fellow, one of eight globally. PACT at UPenn this summer. Wolfram Emerging Leaders Program in the fall. Recognised at ICYS 2026, INSEF India, and the Genius Olympiad.

Questions I Think About

What structural properties of neural networks enable generalisation beyond their training distribution?

Can formal proof systems serve as reliable verifiers for machine-generated mathematical reasoning?

How does entanglement entropy constrain the classical simulability of quantum algorithms?

What are the fundamental failure modes of foundation models applied to biological data?

Can a Hermitian operator be constructed whose eigenvalues correspond to the imaginary parts of the Riemann zeta zeros?

Areas

Machine Learning Theory

Optimisation

Foundation Models

Computational Biology

Quantum Computing

Mathematical Reasoning

Formal Methods

Data Visualisation