AI Data Infrastructure

Every Capability
Covered.
Nothing
Duplicated.

能力维度全覆盖
训练语料零冗余

Dancing Links LLC builds training and evaluation datasets for AI teams. Curated, deduplicated, decontaminated, and versioned — so every run is reproducible and you know what changed.

Scroll to explore

§ 01
— 01 / The Method
Why
Dancing
Links?

"Cover Everything. Repeat Nothing."

Dancing Links is Donald Knuth's technique for exact cover: choose a set of rows that covers every column exactly once — no gaps, no overlap. Pull a node out in constant time, explore, and splice it back perfectly when the branch fails.

That is the training-data problem, exactly. Every capability you care about is a column. Every candidate corpus is a row. The work is covering all of them with as little redundancy as you can afford — and being able to pull a batch back out when it turns out to hurt.

cover(c)   L[R[x]] ← L[x]   R[L[x]] ← R[x]
uncover(c)   L[R[x]] ← x    R[L[x]] ← x

§ 02
— 02 / What We Build
Datasets, Built
Like Software.
01
🧬
Dataset
Curation
Sourcing · Dedup · Decontamination

Sourcing, cleaning, near-duplicate removal, and eval-set decontamination. You get a corpus with known provenance and a written record of what was dropped, not a scraped pile.

02
🏷️
Labeling
& Evals
Rubrics · Agreement · Golden Sets

Human-plus-model annotation with written rubrics, inter-annotator agreement tracking, and held-out golden sets that catch quality drift before it reaches a training run.

03
Data
Pipelines
Versioned · Lineage · Reproducible

Versioned datasets with full lineage, delivered into your training environment. Diff two runs and say exactly which batch moved the number.


§ 03
— 03 / Contact
Contact Me

Business inquiries, partnerships, and collaborations — reach me directly:

Company
Dancing Links LLC
Name
Kaiyuan Feng
Phone
Email

Typically replies within 24–48 hours.