← All work

ML Research / 2026

Ouroboros

Independent research. Deliberate experiments.

I develop language-model experiments in PyTorch and compare proposed mechanisms against explicit baselines.

PythonPyTorchML researchEvaluation

Research — In development

CONTRIBUTION & EVIDENCE
Run the workflow

Follow a research decision.

A high-level research workflow. Model internals, implementation details, and unpublished results are intentionally omitted. Click a step, connection, or trace to inspect it.

Execution trace / Click to inspect

Start the walkthrough, or inspect any step before running it.

Simplified from the project architecture. Runs locally; no live models, credentials, or external services.

Ouroboros / Inspector

Workflow overview

Question

Define the question and what evidence would support it.

Why it matters

Only the public research process is shown. Unpublished architecture, experimental methods, training details, and measurements remain private.

Synthetic payloads explain the boundaries. They are illustrative examples, not project API schemas or live production traces.

The engineering

Inside
the decisions.

Separate a useful model mechanism from an attractive hypothesis.

A proposed model mechanism is only useful if its benefit can be separated from a stronger baseline, extra compute, or experimental noise.

01

The design decision

Test mechanisms individually against explicit baselines before combining them into a larger architecture.

02

The trade-off

Local experiments make iteration possible within hardware constraints. They establish evidence at prototype scale; larger-scale behaviour still needs separate validation.

03

Experiments under local hardware constraints

Develop language-model research prototypes in PyTorch under local hardware constraints.

04

Comparisons against explicit baselines

Compare experimental ideas against explicit baselines and record the conditions under which they are evaluated.

05

Recording limitations and negative findings

Document limitations and negative findings alongside promising results; keep unpublished methods and implementation details private.

Read the system flow

Ouroboros · Public explanation of the implementation boundaries.

  1. QuestionDefine what a proposed change should improve.
  2. BaselineMake the comparison and its resource constraints explicit.
  3. ExperimentRecord the conditions and investigate alternate explanations.
  4. Research notesDocument limitations and negative findings alongside observations.

Current research output

The public artifact is a research log. Experimental work is in development; unpublished methods, architecture, and training artifacts remain private.

Public research log ↗

How to read the exhibit

The walkthrough explains the research process. The optional 3D study is an abstract visual interpretation, with no claim about model architecture or measured performance.

Research — In development

Inspect
the work.

Active independent research with a public research log. This portfolio presents the experimental process; unpublished architecture, implementation, and training artifacts remain private.

Next project: Sylph ↗

Source material

Public research log
Public research log. Unpublished architecture, methods, and training artifacts are not disclosed here.

Opening the project index…

Browse all work ↗

Interactive exhibit