Skip to main content
Research & labs

An eval loop withno invoice attached

Evaluation sweeps, fine-tunes and transcription for unpublished work — reproducible because nothing changes under you.

Novel work leaks by inference too. No cloud trail, nothing to subpoena, nothing to train on.

Overnight batches$8–13k/yr on the meter · here €0

The day, step by step

Each step names the tool and the model doing the work — all from the verified catalog.

  1. Design

    Pin everything

    Weights, quant and prompts pinned in the harness config — the run that works in March replicates in June, because no provider can deprecate your model.

    llama.cpp · pinned GGUF
  2. Nightly

    Sweep

    The evaluation matrix runs overnight at zero marginal cost — the experiments that cloud pricing quietly discourages become routine.

    Custom harness · gpt-oss-120b / Qwen3.5
  3. Weekly

    Fine-tune pass

    LoRA adaptation on the lab's material, evaluated against the pinned baseline the same night.

    Local fine-tune tooling
  4. Ongoing

    Interviews transcribed

    Field recordings and lectures transcribed offline — participant consent covers your machine, not a vendor's.

    Whisper large-v3

The stack, pre-installed

  • gpt-oss-120b
  • Qwen3.5 122B
  • llama.cpp
  • Whisper large-v3
  • LoRA tooling

The honest caveat

One box fine-tunes adapter-scale, not foundation-scale — we say so before you buy. For full-parameter training you still want rented big iron; for evaluation and adaptation, this is the economical shape.

This arrives working — models loaded, agents wired, your team walked through it.

Configure yours