Projects

These projects are the larger threads that organize my work across multiple papers. The publication record lives separately; this page is meant to show the research directions at a glance.

AI for Systems

LLMs, agents, and hardware-aware parallel code

I treat code itself as a computational object for AI: something to read, translate, profile, optimize, and eventually generate in ways that stay grounded in performance and hardware constraints.

Automatic parallelization and translation across OpenMP, MPI, CUDA, and related settings
Domain-specific language models for HPC code and tasks
Performance reasoning, profiling, and hardware-aware code generation

AI for Science

Scientific textures and microstructure

This thread focuses on scientific imagery where texture, morphology, and spatial structure matter. The goal is to build AI tools that understand those patterns rather than wash them out.

Texture-aware foundation models and segmentation
Quantitative metallography and microstructure analysis
Pattern-sensitive scientific imaging pipelines

AI for Science

Experimental physics and feedback-driven processes

Here I work on scientific settings where AI is part of a loop between measurement, physical state, and decision making, especially when experiments or simulations need to be steered in real time.

Multi-view scientific imaging and explainability for experimental setups
Optimization and control in simulation-driven physical processes
Learning systems that close the loop between physical state and computational choice

AI for Science

AI for structural biology

This is the biology-facing side of my AI for Science work, where I collaborate across biological problems that benefit from sequence modeling, language-model adaptation, and scientifically informed prediction.

Language models for RNA editing prediction
Biology-facing modeling built from AI expertise rather than a single fixed method
Collaboration across biological discovery problems

Foundations

Scientific computing, HPC, and systems foundations

The earlier phase of my work remains foundational: source to source parallelization, simulation acceleration, memory systems, distributed algorithms, and the systems view that made the later AI directions possible.

Source-to-source parallelization and compiler-facing HPC work
Memory systems, performance, and simulation acceleration
Distributed algorithms and systems-oriented computation