Research
Research Focus
Machine learning in finance
We use machine learning where it earns its keep — and are candid about where it simply memorises noise. The goal is durable structure, not a leaderboard score.
What we study
- Feature design grounded in market structure, not data-mining
- Out-of-sample validation and leakage control
- Model interpretability and failure modes
- When a simpler model is the honest answer
How we approach it
Overfitting guardrails
Strict train/validation separation and out-of-sample discipline.
Interpretability
We prefer models whose behaviour we can explain and defend.
Skeptical by default
A result must survive costs, regimes and time to count.
This page describes our research discipline and is educational in nature. It is not investment advice, a recommendation, or an offer, and contains no performance figures or predictions.
Other focus areas
Systematic equities
Cross-sectional and time-series signals across the NSE universe.
Options & market microstructure
Order flow, auction theory and the mechanics of how prices form.
Factor research
Value, momentum, quality and low-volatility, tested honestly.
Execution & capacity
Slippage, turnover and the real limits of a strategy at size.
Regime & risk
Drawdown behaviour and stability across full market cycles.
Research dialogue
Discuss a research question with our desk.
Whether you are scoping a mandate or comparing approaches, we are glad to talk through the method.