Research Problem Statement
The complete formulation of the RL-for-trading problem on Indian equities: MDP definition, market context, key challenges, and evaluation protocol. Everything you need to understand what we're solving and why it's hard.
Methodology
High-level overview of our approach: why actor-critic RL, why walk-forward validation, why dual JAX/PyTorch backends, and how we model NSE-realistic costs. The "what" and "why" without the "how" in code.
Open Research Problems
Twenty fundamental questions we have not solved yet. From regime-aware meta-learning to continuous target-weight action spaces. If you're working on any of these, we should talk.
Literature Review
Key papers in RL for trading, meta-learning for finance, Indian market microstructure, and risk-sensitive RL — with our commentary on relevance to NSE-specific problems.