● Proven at a literal terabyte — synthetic, PHI-free

Connect the data the VA already has.

The answers are already in your systems — they just can't agree on who's who. We resolve the same veteran, facility, and contract across dozens of disconnected repositories, and surface the links between them — without moving or replacing anything. Then we proved it end-to-end on a literal terabyte.

See the partnership → See the terabyte proof No-cost CRADA · inside your boundary · data never leaves
1.02 TB
resolved end-to-end
3.95 B
records
94.3%
accuracy (F1), flat
<3%
wrong-merge rate
The problem — in the VA's own words

You bought the platforms. The value is trapped in silos.

The VA has accumulated the data and the systems — but decisions wait, because the same veteran, facility, and contract look different in every repository, and nothing connects across them.

"Don't build data silos. Build enterprise data with bridges across, so it can be used across the agency."— the enterprise-data mandate we keep hearing

Every vendor shows a beautiful front end. The hard part — the part nobody demos — is making the backend actually work the data: resolving identity across systems that share no common key, and proving each connection is right.

What it is

A data fabric — a map over your systems, not another database.

Platform-agnostic. It sits on top of whatever you run — Oracle, SharePoint, cloud, mainframe — and deploys inside your security boundary. We don't migrate your data; we resolve across it where it lives.

Resolve identity across silos

Connect the same veteran / facility / provider / contract across systems that share no common ID — and prove every connection against evidence.

Correlate across domains

Which contracts touch a facility, its operational status, whether that tracks with care delivery — the enterprise question you can't answer today.

Data stays in place

Only the resolved links persist. No central copy, no rip-and-replace, no proprietary lock-in — open, standards-based, in-boundary.

Governed & auditable by design

Every resolution is inspectable and human-reviewable — the foundation a Chief AI Officer and data-governance role can build on.

The proof — measured, not asserted

We resolved a literal terabyte. The accuracy never moved.

The question any data person asks is "does it hold at scale?" So we didn't promise — we ran it: 3.95 billion records across 5,329 independent partitions, each scored against a known answer key. If accuracy degraded with scale, this band would spread. Across a full terabyte, it didn't.

94.31% mean F1
93.5%
94.03 low
94.60 high
95.0%
5,329 partitions · 3.95 B records · standard deviation 0.081 · every score inside a ½-point band. Flat is the result.
1.02 TB
estate resolved end-to-end
94.3% F1
± 0.08 across the full TB
2.97%
wrong-merge — the safe error
linear
throughput scales with cores

Figures demonstrate the method and its scale-invariance on a synthetic, PHI-free, VA-shaped data estate scored against ground truth — not a guaranteed accuracy number on live data. The mechanism holds on real data inside your boundary; that's what a CRADA validates. Run: one laptop, 19.5 hours, partitioned + resolved in parallel — the way distributed entity resolution actually works.

How it works — in plain terms

Weigh the evidence. Only compare what could match. Partition to scale.

Accuracy

Weigh the evidence

Every clue is weighted by how telling it is — a matching SSN is proof, a matching city is coincidence — and the system learns the weights from your data. The same statistical method national census bureaus use to link records.

Speed

Only compare plausible pairs

Comparing every record to every other is impossible at scale. We only compare records that could plausibly match — so it stays fast from a laptop to a terabyte.

Scale

Partition & parallelize

Resolve bounded partitions in parallel — accuracy constant per partition, throughput linear in cores. A laptop proves it; a server does a terabyte in about an hour.

Why us

We sell the answer — on top of what you already own.

The heavyweight platforms

IBM Data Fabric, Palantir Foundry, Snowflake — platforms you buy, marry, and migrate into. Big, slow, lock-in. Another store to move your data out of its systems and into theirs.

Metrics' fabric

Platform-agnostic, in-boundary, no lock-in — and we prove the hard part (resolution + correlation) on your data, fast. We can even ride on top of the platform you already bought. No rip-and-replace.

The partnership

A no-cost CRADA — prove it on VA data, inside VA's boundary.

We're not asking the VA to buy anything or migrate anything. Through a Cooperative R&D Agreement via VA Pathfinder, we co-develop and validate the fabric on your data, in your environment — no funds exchanged, data never leaves.

Metrics provides

The data fabric, the engineering team, and the build — co-invested, at no cost to the VA.

The VA provides

Data access, environment, and domain expertise — inside the VA boundary; data never leaves.

1 · CDA

A short confidentiality agreement to scope feasibility and share safely.

2 · PII-CRADA

Co-develop & validate — synthetic → de-identified → real, under protections.

3 · Field it

Prove on real data; ready the enterprise rollout.

Prove a cross-silo capability you can't buy off the shelf.

On your own data, inside your boundary, with zero acquisition risk. The next step is a short conversation.

Email Phill Sieg →
Phill Sieg · Business Development, Metrics, LLC · psieg@metrics-llc.com