I'm a software engineer interested in GPU computing, numerical software, HPC, and performance-oriented systems.

I currently work professionally as a full-stack engineer, while my personal work focuses increasingly on Rust, GPU execution, numerical computing, and heterogeneous systems.

Before moving into software engineering, I worked in mechanical / structural engineering and used ANSYS and Abaqus for finite-element analysis. That background is still a major influence on the kinds of software problems I enjoy exploring: numerical methods, solvers, parallel computation, and the systems underneath them.

A hands-on exploration of GPU-oriented systems programming in Rust, progressing from low-level interoperability to a small heterogeneous compute framework.

The project covers:

• Rust ↔ C / C++ FFI
• Vulkan compute with ash
• explicit GPU resource ownership and synchronization
• GPU memory pooling and suballocation
• device-local memory and staging transfers
• compute graphs and dependency scheduling
• multi-workgroup reductions and numerical GPU kernels
• CPU / GPU execution backends
• JSON-driven mini compute framework
• ARM64 cross-compilation
• Vulkan workloads on NVIDIA Jetson Orin Nano via Slurm

The lab intentionally builds these layers explicitly rather than hiding them behind a high-level GPU framework.

A browser-based finite-element–style application built from scratch using Rust, WebAssembly, WebGL, and WebGPU.

It explores the pipeline from engineering data structures to numerical execution and visualization:

• FEM-style assembly
• sparse linear algebra
• CG / PCG iterative solvers
• Block-Jacobi preconditioning
• GPU compute kernels and reductions
• WebGPU computation
• WebGL visualization
• Rust compiled to WebAssembly

Related experiments include wgpu_solver_backend and wgpu_solver_slurm, extending the solver path from browser GPU execution toward native and scheduled workloads.

An experimental infrastructure lab for understanding how AI workloads fit into heterogeneous compute environments.

It combines experiments around:

• Rust-based tooling
• Slurm scheduling
• K3s
• RAG pipelines
• llama.cpp inference
• LoRA fine-tuning
• shared storage and GPU workloads
• Raspberry Pi + NVIDIA Jetson hardware

The focus is on understanding the infrastructure layers rather than building a production AI platform.

My main technical interests are:

• GPU computing — Vulkan, WebGPU, compute shaders
• Numerical computing — sparse linear algebra, iterative solvers, FEM
• HPC and heterogeneous execution — scheduling, GPU workloads, distributed compute infrastructure
• Rust and systems programming
• WebAssembly and browser-native compute
• building abstractions without losing sight of what happens underneath them

I particularly enjoy projects that connect multiple layers:

algorithms → memory → GPU execution → runtime → infrastructure

Languages: Rust · TypeScript · JavaScript · Python

GPU / Compute: Vulkan · ash · WebGPU · WGSL · WebGL · wgpu

Systems / HPC: Slurm · Linux · Ansible · NFS · Apptainer · ARM64 · NVIDIA Jetson

Web / Backend: Angular · Vue · NestJS · Node.js · Redis · Kafka · Kubernetes

Numerical: sparse matrices · CG / PCG · Jacobi / Block-Jacobi · FEM concepts

I'm currently especially interested in opportunities around Rust, GPU computing, numerical software, HPC, and performance-oriented engineering.