Portfolio

Optimal Transport in Linear ICA

Master Thesis @ MPI for Intelligent Systems and University of Tübingen

Ongoing Research into applying optimal transport theory to linear Independent Component Analysis (ICA) prolem. Git repository access provided on request.

Supervisors: Dr. Simon Buchholz, Prof. Dr. Michel Besserve, Prof. Dr. Joachim Grammig

NICE Toolbox (Computer Vision & DL)

Graduate Research Assistant @ MPI for Intelligent Systems

Core contributor to the Nonverbal Interpersonal Communication Exploration (NICE) Toolbox. Implemented Deep Learning models for Pose Estimation, Emotion Detection, and Head Orientation. Designed the asset manager for all algorithms network weights and structured the testing and Docker containerization pipelines.

Structured Product Design, Pricing & Hedging (Financial Engineering)

KU Leuven (Course: Financial Engineering)

Designed, priced, and hedged a Bonus Certificate linked to Costco (COST). Used the Bates Model (stochastic volatility + jumps) with a two-stage calibration and Monte Carlo simulation for path-dependent payoff pricing.

Probabilistic Asset Pricing

Master Seminar @ University of Tübingen

Developed a Bayesian framework for estimating stock risk premia. Integrated the hybrid model of Grammig et al. (2024) with forward-looking measures from Martin & Wagner (2019) using Kalman Filtering and the EM algorithm.

Supervisor: Prof. Dr. Joachim Grammig

A Comparative Study of CBM and UCCAPM

University of Tübingen (Empirical Asset Pricing)

Comparative analysis of Consumption-Based (CBM) and Ultimate Consumption (UCCAPM) models on 25 Fama-French portfolios. Investigated model performance during COVID-19 and the impact of lagged consumption adjustment.

ARMA Process Analysis & Estimation (Time Series Analysis)

University of Tübingen (Advanced Time Series Analysis)

Detailed analysis of Conditional Maximum Likelihood (CML) vs Quasi-Maximum Likelihood (QML). Performed simulation studies to test stationarity, efficiency, and robustness of confidence intervals.

Ripple Down Rules Simulation (Tree-based Algorithm)

BITS Pilani (APPCAIR Lab)

Implemented a Python simulation for Ripple Down Rules (RDR), a tree-based incremental learning approach.

Supervisor: Prof. Ashwin Srinivasan