Markets are complex
adaptive systems.
Traditional forecasting assumes markets behave like clocks: deterministic, linear, predictable. They don't. They behave like ecosystems: regime shifts, feedback loops, emergent phenomena. We build models that adapt.
We don't predict direction.
We predict magnitude.
Most quantitative strategies try to forecast whether a stock will go up or down. We ask a fundamentally different question: which stocks will experience outsized moves this week, regardless of direction?
We engineer systematic, non-directional options structures that capture alpha from volatility itself. Hedged. Defined-loss. The model adapts across market regimes (calm, mixed, or stressed), reconfiguring signals dynamically.
This isn't a black box. It's a research-driven, walk-forward validated system built on institutional-grade data infrastructure and rigorous statistical methodology.
Fully systematic. Data-driven.
Zero manual intervention.
Signal Discovery
We ingest and process data from 44 distinct sources spanning derivatives, equities, volatility surfaces, institutional flows, macro indicators, and alternative datasets. Every signal is computed with strict no-lookahead discipline.
Regime Detection
Our adaptive system identifies whether the market is in a calm, transitional, or stressed regime and dynamically reconfigures the model accordingly. Different market states demand different signals.
Prediction Engine
Walk-forward machine learning ensemble with rolling training windows and weekly retraining. No overfitting. No curve-fitting. Out-of-sample validation across every prediction.
Execution
Non-directional options structures that capture predicted magnitude moves. Hedged, defined-risk positions. Direction is irrelevant. We trade the size of the move, not the direction.
Options pricing models.
Two complementary models power our pricing and IV extraction pipeline: Black-76 for futures-settled derivatives and Black-Scholes as the universal fallback.
Black-76
PRIMARYThe primary model for NSE futures-settled options. Standard for Indian derivatives markets where all options are European-style and cash-settled against the futures price. Uses the forward price directly, eliminating dividend and borrow-rate assumptions.
d₁ = [ln(F/K) + 0.5σ²T] / (σ√T)
d₂ = d₁ − σ√T
Call = e−rT [F·N(d₁) − K·N(d₂)]
Put = e−rT [K·N(−d₂) − F·N(−d₁)]
F : Forward price (3-way: futures mid → parity → spot fallback)
K: Strike price | σ: Implied volatility | T : Time to expiry
Black-Scholes
FALLBACKThe classic spot-based model. Used when futures data is unavailable. Includes continuous dividend yield adjustment for accurate equity option pricing.
d₁ = [ln(S/K) + (r − q + 0.5σ²)T] / (σ√T)
d₂ = d₁ − σ√T
Call = Se−qTN(d₁) − Ke−rTN(d₂)
Put = Ke−rTN(−d₂) − Se−qTN(−d₁)
S: Spot price | q : Continuous dividend yield
r: Risk-free rate | N(·) : Cumulative normal distribution
IV is extracted via Brent's method (Brentq), a hybrid bisection/secant root-finding algorithm with superlinear convergence. Each option in the chain is individually solved, building the full IV surface from observed market prices.
Full analytical Greeks computed at every strike in the chain, enabling position-level risk decomposition and hedging.
Quote Hygiene System
Two-tier filtering ensures only institutional-quality quotes enter the pricing pipeline. Strict tier first, lenient fallback if insufficient depth.
IV Smile Calibration
Four stickiness models capture how the volatility surface evolves as the underlying moves, critical for accurate scenario-based P&L.
GPU-accelerated combo tester.
Our proprietary engine evaluates millions of multi-leg options combinations using GPU-accelerated matrix operations, finding the optimal structure for every prediction signal in seconds.
Enumerate
Generate all valid multi-leg combinations from the live option chain, respecting strike range, expiry, and type constraints.
Price
Re-price every combination across a ±10% spot scenario grid using calibrated IV smile and the Black-76 model.
Filter
Reject any structure that violates risk constraints: defined maximum loss, profitability across ±3-10% moves, no naked exposure.
Rank
Score surviving strategies by risk-adjusted return on capital, selecting the structure with the best defined-risk payoff profile.
VIT-TBI
VIT Technology Business Incubator
VIT University, Vellore
SIIF
SSCBS Innovation & Incubation Foundation
Shaheed Sukhdev College, Delhi University