Jul 2023 – May 2025 · 1 yr 10 mos · Hyderabad, Telangana, India · On-site

• Quant Research:
• > Played a key role in developing ICE's cutting-edge margin model IRM2 (https://www.ice.com/clearing/margin-models/irm-2), utilizing a filtered historical simulation and a non-parametric Value at Risk (VaR) approach to calculate margin requirements across the entire ICE futures and options derivatives complex.
• > Developed volatility surface fitting methodologies to adaptively model assets across the liquidity spectrum, from illiquid to highly liquid. These methodologies enhanced the exchange’s settlement process by promoting fair pricing and mitigating spread, butterfly, and calendar arbitrages.
• > Designed and implemented a comprehensive quantitative backtesting framework, incorporating advanced algorithms for linear and nonlinear risk factor construction, as well as pricing models such as Black-Scholes, Bachelier, and moment matching.
• > Performed time series analysis to improve model volatility estimation, addressing key factors like volatility clustering, procyclicality, and seasonality. Additionally, analyzed model risk sensitivity during market stress, focusing on volatility estimation and correlation risk in well-hedged portfolios.
• Software Development:
• > Developed libraries for various pricing functions and statistical models in C++, leveraging Pimpl (Pointer to Implementation, C++ design idiom) to enhance stability and scalability for production use.
• > Utilized Boost, Eigen, Math, Thread, and Pointer libraries to ensure high-quality software development, addressing feature enhancements and bug fixes.
• > Worked on database migrations using PostgreSQL, creating relational database structures for an internal backtesting system, and wrote multiple Python scripts for automation.
• Leadership:
• > Pioneered recruitment for the Indian team as the first member, overseeing the hiring process, conducting 100+ phone interviews, and successfully recruiting three candidates.

May 2022 – Jun 2023 · 1 yr 1 mo · Greater Bengaluru Area

• Paper: https://arxiv.org/abs/2405.01040
• > Introduced an innovative FSCIL (Few-Shot Class-Incremental Learning) framework that incorporates a language regularizer and a subspace regularizer to enhance model performance and flexibility.
• > Utilized OpenAI’s CLIP (Contrastive Language–Image Pretraining) architecture to map image features into a semantic space defined by class labels, enhancing the accuracy and relevance of image classification tasks.
• > Developed novel loss functions to enhance the base model's robustness, specifically targeting a reduction in forgetting and improving model retention over time.

Jun 2021 – Jul 2021 · 1 mo · Bangalore Urban, Karnataka, India

• > Analyzed Customer Relationship Management (CRM) metadata from approximately 1 million business transactions to optimize the search for potential customers, improving efficiency and effectiveness in targeting prospects.
• >Conducted regression analysis and designed data visualization tools to generate actionable insights, providing feedback to the business team to enhance operational efficiency.

May 2020 – Jul 2020 · 2 mos · Mumbai, Maharashtra, India

• Developed a Stock Portfolio Tracking System in Python to assess the daily performance and progress of the portfolio, enabling detailed evaluations over time.