AI Engineering

By Stephen Meier

AI Engineering

Before language models, I simulated excitable media like the waves shown here. The habits carried over: build the model, measure carefully, report what you find. These are my notes on fine-tuning and evaluating LLMs.

Four snapshots of a simulated two-dimensional excitable medium. Red wave fronts on a blue ground fold into four-fold symmetric shapes: a scalloped ring, a cross with a bright centre, crescents moving apart, and the ring again.
Excitation waves in a two-dimensional reaction-diffusion system in the soliton-like regime, at T = 0, 5.5, 11.5 and 17.5 ms. Re-rendered from the simulation data for Fig. 32 of my dissertation (2016).
  • Timing Agentic RL for Qwen3.8-27B on Two DGX Sparks

    How long would an agent RL run with Qwen3.8-27B take on two DGX Sparks? I measured generation and training separately, tested speed-ups one at a time, and then traced why long agent sessions slowed to 2 tokens per second.

    Read the post

  • Can an Open-Weight Model Prove Code Never Crashes?

    Anthropic had Claude prove that real programs cannot crash. I tried the same loop at small scale with an open-weight model on two DGX Sparks: GLM-5.3-Flash writing Verus proofs for eight Rust routines. It proved all of them crash-free and found every planted bug, but one fix silenced a crash without fixing it, and the…

    Read the post

  • Async GRPO on Two DGX Sparks: Same Curve, Different Bottleneck

    Hugging Face trained Qwen2.5-Math-1.5B with TRL’s asynchronous GRPO trainer on four H200s. I reran it on two DGX Sparks on my desk. The reward curve reproduced; the tuning advice did not, and one of the speed knobs quietly changed the training objective.

    Read the post

  • 2019PD-1/PD-L1 immune checkpoint and p53 loss facilitate tumor progression in activated B-cell diffuse large B-cell lymphomas. Blood. doi:10.1182/blood.2018889931
  • 2018The cytidine deaminase under-representation reporter (CDUR) as a tool to study evolution of sequences under deaminase mutational pressure. BMC Bioinformatics. doi:10.1186/s12859-018-2161-y
  • 2016The relationship between Nernst equilibrium variability and the multifractality of interspike intervals in the hippocampus. Journal of Computational Neuroscience. doi:10.1007/s10827-016-0633-5
  • 2016Non-invasive evaluation of cardiac repolarization in mice exposed to single-wall carbon nanotubes and ceria nanoparticles via intratracheal instillation. Environmental Science: Nano. doi:10.1039/c5en00225g
  • 2015Bursting regimes in a reaction-diffusion system with action potential-dependent equilibrium. PLOS ONE. doi:10.1371/journal.pone.0122401
  • 2012Multiple mating but not recombination causes quantitative increase in offspring genetic diversity for varying genetic architectures. PLOS ONE. doi:10.1371/journal.pone.0047220