Abhishek Choudhary

Experience

• 

  PwC India

  Jul 2017 - Sept 2019

  Consultant

  My responsibilities as a cybersecurity consultant included maintaining Endpoint security solutions and conducting research on Blockchain technology case studies.

• 

  ZS

  Mar 2021 - Jul 2021

  Data Science Associate - Intern

  I worked for a pharmaceutical client to forecast the patients’ enrollment duration time for a rare disease (HDFN) in phase III of the clinical trial using simulation techniques (monte Carlo Markov chain and method of moments). Forecasting of patients’ enrollment is done by fitting a suitable distribution (Poisson-gamma distribution).

• 

  EY

  Aug 2021 - Jul 2022

  Consultant

  Worked on model validation for a CCAR model for a Commercial Real Estate portfolio and onend-to-end validation of a scorecard model for a wholesale credit risk portfolio.In addition, I was also involved in research on the interpretability of trusted AI and machine learning models (group fairness, individual fairness, counterfactual, and AIA).

• 

  Citi

  Jul 2022 - now

  Assistant Manager

  Credit Risk: : Validated Probability of Default (PD) models for wholesale portfolios (Corporate, Sovereign, and Bank Exposures) and Retail Portfolios (Mortgages, Revolving, and Non-Revolving Credit) for local regulatory, ICAAP, and IFRS 9 reporting purposes, using both standardized and internal rating-based approaches (IRB). Methodologies include logistic regression, Gaussian copula (bi-variate / trivariate), Vasicek loan portfolio models, and Merton models.Validated an LGD model that transforms ORR data into the FRR distribution using polynomial regression, exponential transformations, and Beta distribution to derive the LGD curve, allowing downstream models to generate stress LGD scenarios for risk capital calculation via Monte Carlo simulation.Market Risk: : Conducted end-to-end model validation for the RWA calculation of the market risk component for ICAAP purposes. The RWA computation includes: GIRR and currency risk, aligned with the FRTB guidelines. The methodology involved linear regression and ARIMAX techniques. Also worked on some validation activities (OPA and AMR) for market risk models (based on the FRTB and IRRBB guidelines) to compute RWA and VaR for the portfolio within the scope of the model.Derivatives Pricing: : Worked on end-to-end model validation for a trading strategy that can be used for share repurchase during the blackout period, in line with regulatory requirements. The underlying share price evolves via geometric Brownian motion and Monte Carlo simulation. Show less