‍Live - Prediction Markets Statistical Arbitrage Engine - Polymarket
Live statistical arbitrage strategy trading pricing inefficiencies in crypto prediction markets using options-implied probabilities and structural arbitrage.

What actually mattered
- Probability first: system attempts to quantify the probability of the event rather than trade purely on momentum.
- Trade the discrepancy: The objective is to identify P_model ≠ P_market with enough margin to justify taking risk.
- Exploit structural relationships: Related prediction contracts can contain arbitrage opportunities even without forecasting the underlying asset.
- Price execution matters: strategy evaluates executable bid/ask prices rather than relying exclusively on midpoints.
- Fees are part of the signal: An edge that disappears after fees is not an edge.

Systems that model order flow, liquidity, and matching dynamics in exchange-style environments.
Focused on market microstructure, execution dynamics, and probabilistic decision-making in short-horizon trading systems.
‍
Live - Market Making Trading System - Adaptive Liquidity Provision System
Live market-making engine combining microstructure alpha, fair value, toxicity, and execution-aware quoting under realistic market conditions for systematic liquidity provision.
‍
What actually mattered
- Alpha is horizon-dependent: current micro signal alpha exhibits increasing predictive information with horizon
- Regimes determine participation: the same signal can behave differently across market states.
- Toxicity is fill-dependent: the key question is whether a fill represents favorable liquidity provision or adverse selection.
- Execution drives realized edge: queue position, latency, snapshot age, and cancellations can outweigh marginal alpha improvements.

‍Market Microstructure Model - Limit Order Book Simulation
High-performance C++ limit order book simulating price-time priority execution, order flow, and liquidity dynamics.
‍
What actually mattered
- Queue position is a primary driver of execution probability in passive order flow
- Displayed liquidity is not equivalent to executable liquidity under fast order flow
- Execution outcomes are path-dependent and sensitive to order arrival sequence
- Liquidity consumption dynamics determine short-horizon fill uncertainty

Monte Carlo-based frameworks for evaluating decisions under uncertainty and partial information.
‍
Monte Carlo Decision Engine - Adversarial Sequential Decision Simulator
Multi-agent simulation engine for probabilistic decision-making under hidden information and dynamic state evolution.

What actually mattered
- Decision quality depends on outcome distribution, not just EV
- Risk-adjusted evaluation improves robustness under uncertainty
- Sequential dependency amplifies exposure to state uncertainty
- Optimal actions are sensitive to changes in underlying assumptions

Pricing, hedging, and volatility-driven portfolio construction under stochastic market conditions.
‍
Delta Hedging & Volatility Trading Simulator
implementing dynamic delta replication, risk tracking, and PnL decomposition under discrete-time rebalancing.

What actually mattered
- Discrete hedging creates unavoidable replication error
- Gamma dominates PnL near expiry and during large moves
- Volatility mis-specification is the main driver of PnL deviation
- Option PnL is path-dependent rather than mark-to-model static

Systems for automated market data ingestion, transformation, and execution-style simulation.
‍
Market Data Pipeline
Automated market data ingestion pipeline for XML financial data, with ETL processing and storage in Azure SQL for trading system integration.

What actually mattered
- Position sizing was constrained more by risk / margin than pricing output alone
- Exposure sensitivity mattered more than absolute model accuracy
- Small regime changes in volatility significantly altered portfolio risk
- Consistency between pricing and risk systems was critical for stability

Market Data Parser & Transformation Pipeline
XML market data parser for transforming hierarchical financial datasets into structured inputs for quantitative trading systems.

What actually mattered
- Data consistency is critical for trading correctness
- Small schema inconsistencies amplify into PnL distortion
- Most model failures originate from data, not modeling
- Market data is a first-order driver of strategy quality

Systems modeling and control projects in robotics and embedded systems, focused on dynamic system behavior and constrained optimization.
- Control systems (multi-actuator dynamics, feedback control)
- Embedded systems (autonomous actuation, real-time control)
‍

Continuously Variable Transmission (Dual-Motor Control System)
Dual-motor system modelling variable torque transmission dynamics.

Smart Trolley Locking & Access Control System
App-controlled locking system for secure retail automation, integrating authentication, locking mechanisms, and real-time state tracking through an app-based interface.

Autonomous Delivery Trailer (Embedded Control System)
Autonomous delivery trailer with embedded sensing and control logic to enable stable motion and responsive actuation in a mobile logistics environment.

Custom LEGO Technic - Audi RS6
Custom multi-actuator vehicular control system modelling drivetrain and suspension systems.