I’m a software engineer focused on the intersection of quantitative finance and distributed systems. I specialize in building high-performance Svelte applications that bridge the gap between complex backend logic and intuitive, low-latency user interfaces.
Currently, I am focused on architecting terminal-centric tooling and reactive web systems, notably as the creator of difi and zsweep (HN #5 & #6). Beyond my own projects, I am an Open Source Contributor to Neovim and Statue. I’m also deep-diving into market microstructure—working through The Green Book to apply systems engineering rigor to quantitative finance.
When I’m not at the terminal, you can usually find me at the Muay Thai gym, training calisthenics, or behind a drum kit. I’m an avid off-roader in my Tacoma TRD Pro and a heavy consumer of specialty coffee. I am currently completing my double major in Computer Science and Mathematics at the University of Virginia.
Experience
Research Intern · UVA Biocomplexity Institute
Formulating massive-scale network simulations to model the stochastic spread of viral pathogens. Engineering attention-based Graph Neural Network (GNN) architectures to predict multi-hop epidemic propagation pathways using spatiotemporal message-passing frameworks.
- GNN
- Network Science
- Simulation
- Bayesian Data Science
- HPC
Research Assistant · UVA (BAL Lab)
Formulating a quantum-inspired psychometric framework utilizing Hilbert-space probability models. Engineering a mathematical framework for multivariate longitudinal missing data imputation that enforces the Law of Total Variation and architecting a C++17 Lagrangian projection engine.
- C++
- RcppArmadillo
- Psychometrics
- Optimization
- Mathematical Statistics
Research Assistant · Yale (Liu's Group)
Developing a deep unrolled ADMM framework for inverse problems in computational imaging. Engineering a specialized PyTorch pipeline with custom U-Net architectures and physics-informed constraints for multi-dimensional spatial and spectral data reconstruction.
- PyTorch
- ML
- ADMM
- Computational Imaging
- Physics-Informed Neural Networks
Research Assistant · UVA (Ke's Group)
Applying quantum mechanics principles to engineer complex Density Functional Theory (DFT) reconstruction algorithms. Orchestrating a cloud-native ML pipeline using Go and Next.js and creating immersive browser-based 3D atomic visualizations.
- Go
- Next.js
- DFT
- Three.js
- Quantum Mechanics
- Machine Learning
Projects
smriti — Longitudinal Imputation Engine
An R package for automated longitudinal missing data imputation. Executes a three-phase architecture combining missForest initialization with a custom C++ Lagrangian projection engine to strictly preserve the structural variance of the target covariance manifold.
- R
- C++
- RcppArmadillo
- Statistics
- CRAN
difi — Git Diff Review Tool
A high-performance CLI tool built in Go for interactive Git diff reviews. Features a keyboard-centric Terminal User Interface (TUI) with a file tree and editor-aware navigation, allowing users to jump directly to specific lines in Neovim/Vim for rapid code refinement.
- Go
- Bubble Tea
- Git API
- CLI
- Nvim Plugin
sanjaya — Academic Graph Pipeline
An automated academic literature extraction pipeline utilizing the OpenAlex Academic Graph API, Scrapy, and Playwright. Extracts bilingual data and exports structured CSV/JSON datasets for downstream interdisciplinary research applications.
- Python
- Scrapy
- Playwright
- Data Engineering
- Academic Graph
Neovim — Contributor
Contributed multiple Pull Requests to the Neovim core (C/Lua). Focused on Lua state change and ENV variable config supporting Vim logic.
- C
- Lua
- Open Source
- Systems
Certifications
AWS Certified AI Practitioner · Amazon Web Services
Validated expertise in deploying production-grade AI solutions on AWS. Focused on prompt engineering, fine-tuning Foundation Models via Amazon Bedrock, and implementing low-latency inference pipelines for real-time financial data processing.
AWS Certified Cloud Practitioner · Amazon Web Services
Mastery of the AWS Well-Architected Framework, emphasizing security, high availability, and performance efficiency. Architected cloud-native environments leveraging Amazon Aurora for relational data and EC2/Lambda for scalable compute logic.
USACO Silver Division · USA Computing Olympiad
Competed in high-stakes algorithmic challenges focusing on computational efficiency and data structure optimization. Solved complex problems requiring $O(n log n)$ performance using greedy algorithms, dynamic programming, and graph theory.
Loosely designed in Figma and engineered in Neovim on an HHKB. Built with SvelteKit and Tailwind CSS, reviewed with difi, and deployed with Vercel.