Competitive programming and performance-critical algorithm work.

Primary language for research code, experiments, and tooling.

Shell scripting, environment setup, and batch automation.

Neovim configuration and editor environment scripting.

Formal theorem proving and proof engineering.

Model implementation, training loops, and debugging.

Transformers, Datasets, and Accelerate for model loading and fine-tuning.

Quantization, KV cache, and training pipeline acceleration.

W&B and TensorBoard for run comparison and reproducibility.

High-throughput inference serving and evaluation.

Cluster job submission, scheduling, and management.

Environment packaging and isolation for reproducible experiments.

Development environment, toolchain, and troubleshooting.

Version control and open-source collaboration.

Python environment and dependency management.

Typesetting for papers and reports.