Mumbai, Maharashtra, India
Hi, I’m Amit. I work as an Authorized Officer (Quantitative Analyst) in the Quantitative Risk Modelling division of UBS Investment Banking, Mumbai. My work spans risk factor model development and backtesting (Local Vol, GBM, QGM, PK, Hull-White) across all major asset classes—Equities, FX, IR, Credit, Basis, and Inflation. I also drive AI-focused solutions at UBS, regulatory remediations, and model implementations, ensuring both innovation and compliance in a highly regulated environment. As part of the Exposure Model Performance Crew, I contribute to enhancing model accuracy, stability, and predictive power through advanced risk analytics. Previously, I worked as a Research Analyst at the Reserve Bank of India (RBI). Academically, I hold a B.Sc. (Hons.) in Physics from Jamia Millia Islamia and a MSc with a Minor in Computational Economics from IIT Jodhpur. I qualified GATE in Economics (XH) with an All-India Rank of 189 placing in the top 8%. I was also selected for the PhD program in Quantitative Finance at IIT Madras—a recognition of my research potential—though I later stepped away due to personal challenges. My research interests span financial mathematics, risk management, machine learning, and financial economics. Beyond work, I actively contribute to research and open-source projects, including an R package (CustomDerivative) for exotic derivatives pricing and a Python-based stochastic process simulator (StochPathSim). My research includes an extension of the Hull-White interest rate model, published on arXiv, which improved bond pricing accuracy by 14%. I’m fluent in Python, R, with basic knowledge of C++, and currently exploring Julia. I also have experience with Bloomberg and Reuters Terminals, and working familiarity with Tableau, and Power BI. At my core, I’m passionate about bridging mathematics, finance, and AI to build robust models, solve real-world financial challenges, and push forward the boundaries of risk management.
Model Development Quant 1. SFT/OTC/ETD Risk Factors Model Development & Backtesting Framework: "Developing and refining Equity, Interest Rates, FX, Basis, and Credit risk factor models in ACQA through a streamlined backtesting framework for SFT/OTC/ETD derivatives, enhancing risk precision and ensuring regulatory compliance." 2. Automated Simulation Results Analysis Tool. "Designed and developed an internal automated tool to analyze simulation results for various models under UBS Group and UBS ESE, enabling seamless evaluation of risk factor distributions, statistical measures, and performance metrics, driving informed decision-making. Presentated this to Global QRM head & other senior stakeholders. 3. Exposure Model Monitoring Enhancement: "Advancing SFT Pillar 1 ESE model monitoring by automating KMPI assessments, improving efficiency, accuracy, and scalability in risk management." 4. Developed a module which processes a large dataset ( 1 bn+ rows / 10 GB+ ) into parquet under the limit of Memory+Time constraints 8 GB RAM only inside Jupyter Notebook. 5.Built CodeNarrator using o1 models on UBS infra — enabling ‘Chat with Code’, ‘Chat with Data’, and ‘Chat with PDF’ & ' MR diff Analyzer ' to boost model developer productivity and document understanding by 5x+ across teams.
• I was employed in the Division of International Finance, DEPR. • Major Project: High-Frequency Trading Analysis – Analyzed 3TB+ tick-by-tick data on INR/USD ETFs to study High- Frequency Trading’s impact, providing key inputs for monetary policy adjustments and market stabilization strategies. • Assisted senior analysts in model validation and adjustments, improving macroeconomic and financial models' efficacy by 45% and enhancing policy recommendations. Led time series forecasting research, advancing macroeconomic models to better predict inflation and growth trends, directly influencing policy decisions.
• Developed Generative AI Product for EdTech Company "QuestGPT" to enhance conversation with YouTube Transcription. • Developed GAI WebApp "LegalAI" for making legal personnel life easier. The WebApp caters to Judiciary, IPC, and Company Act. • Successfully built a product for US-based clients, which utilized deep learning to predict student performance. • Developed a dashboard for a client based in Kenya, which offered real-time data on various solar power plants. • Created a Time series forecasting model for a Pharma based industry using ARIMA and exponential smoothing. • Executed a project for PSG, Kenya, to create an automated data dashboard and implementation of AI for the alert system. • Designed and developed ResumeParser using NLP, a resume parser that can extract relevant information from folders of CV/resumes. This tool has made life easier for HRs in organizations. Tools used: Python, Git, Excel, SQL, Machine Learning,LLMs and Time Series.