Hi, I'm Kalash

Proprietary trading firm (NDA), Gurgaon · systematic energy futures desk.

Researched WTI crude calendar spreads across 1-, 3-, 6- and 12-month tenors, through both deep contango and steep backwardation. Split tenors by Hurst exponent into mean-reverting and trending groups, which drove entirely different trade construction, and built session-aware trading windows after finding the same spread behaves differently depending on who is in the market.

Built the desk’s C#/.NET execution gateway bridging Python strategy engines to the exchange via the TT SDK, and a five-layer pre-trade risk engine with an automatic kill-switch on a daily P&L breach. Research was held to purged walk-forward validation with embargo and ranked on deflated Sharpe, with full costs applied before any result counted.

Built to teach myself how the WTI forward curve actually behaves. 421 exchange-quoted calendar spreads across 8 years of 1-minute data: term-structure PCA into level, slope and curvature, a box and butterfly playbook, and a playable market-making game with Glosten–Milgrom adverse selection. A study, not a strategy — no alpha claim.

Market-making simulator. Quote around a jump-diffusion fair value while uninformed flow pays the half-spread and informed flow arrives ahead of jumps. P&L decomposes into spread capture, inventory risk, adverse selection and liquidation — an identity asserted to 1e-6 across 34 tests.

Streamlit backtester trading six butterflies simultaneously along the curve (front+4 through front+9). Thresholds are adjustable live, so the sensitivity of the trade population is visible rather than hidden behind a single backtest.

Cross-sectional mean reversion on crude spreads. A rolling 252-day PCA strips out the common factor and the residual is traded on multi-horizon z-scores, with an Ornstein–Uhlenbeck half-life converting the signal into a holding period.