Research Interests
I work at the intersection of quantitative finance, machine learning, and systems engineering. Here's what I'm currently exploring and thinking about.
Current Focus Areas
Interest Rate Derivatives & Macro Strategy
At J.P. Morgan, I focus on developing and analyzing strategies across interest rate products including swaps, swaptions, swap spreads, and futures.
- •Portfolio scenario analysis across rate hike/cut environments and curve shifts
- •Option pricing using binomial trees for vanilla and EOM options on futures
- •Coverage of Treasuries, TIPS, and short duration strategies
- •Wildcard scenario estimation for tenor-specific treasury baskets
AI & Machine Learning
I'm fascinated by the rapid progress in AI, particularly large language models and their underlying mechanisms.
- •Transformer architectures and attention mechanisms
- •Scaling laws and their implications for AI progress
- •Applications of ML in quantitative finance
- •Transfer learning and fine-tuning strategies
AI Hardware & Infrastructure
Understanding the hardware that powers AI is crucial for understanding its trajectory.
- •GPU architectures (NVIDIA's roadmap, AMD's competition)
- •Custom AI accelerators (TPUs, custom ASICs)
- •Memory bandwidth constraints and solutions
- •Data center infrastructure and energy considerations
High-Performance Systems
Building systems that perform well under load and scale gracefully is both an engineering challenge and an art.
- •Algorithmic optimization for financial data processing
- •ETL systems for high-frequency and daily trading models
- •Python performance optimization (Pandas, NumPy)
- •Modernizing legacy systems (C++ to Python migrations)
Past Research
Applications of AI and IoT in Healthcare
Undergraduate Thesis, BITS Pilani | January - May 2022
Documented the trajectory of AI and IoT in medical diagnostics, focusing on NVIDIA's GPU progress and applications in healthcare.
- •Tested NVIDIA Jetson Nano/Xavier Kits for medical imaging inference
- •Analyzed loss functions, optimizers, and accuracy parameters for medical ML
- •Explored federated learning for privacy-preserving medical AI
Current Questions
Some questions I'm actively thinking about:
"What are the true bottlenecks to AI progress - compute, data, or algorithms?"
"How will AI change the nature of quantitative finance work?"
"What's the right way to think about AI capabilities vs. limitations?"
"How do you build systems that remain maintainable as they grow?"
I occasionally write about these topics on my blog. I'm also always happy to discuss ideas with people working on similar problems.