ΒΑΣΑΝΟΣ · n. touchstone · agent reliability
The stone that reads the streak.
Basanos hardens AI agents for companies that can’t afford to guess — eval audits that block bad ships, judges that survive their own audit, and post-training you own.
34%
Fewer false positives — fraud ML, Capital One
850K/hr
Exactly-once event pipeline — Capital One
10M+
Daily LLM inferences — IBM
+25%
Domain-task lift from SFT — IBM
Real numbers from the founder’s decade in regulated systems. The track record →
The streak test
Would your judge survive its own audit?
Most agents ship on vibes: a demo went well, a spot-check looked fine, nobody saw the failure modes coming. In a regulated workflow, a wrong answer isn’t embarrassing — it’s exposure.
Evals fix that, but only as far as the judge is honest. An LLM judge that misses violations — or hands out passes for the right magic words — turns your safety net into a rubber stamp.
And the moment scores steer training instead of dashboards, the stakes escalate: would your reward survive RL pressure?
The ladder
Each rung is built from the one below it.
The whole circuit
One judge, one closed loop.
Every engagement feeds the next: the harness scores the agent, the scores become the archive, the archive hardens the judge, the judge trains the model — and the trained model comes back as the agent, certified on the same harness that started the loop.
THE AGENT
runs the live workflow
THE HARNESS
three judges gate every ship
SCORED TRACES
the archive grows
HARDENED JUDGE
survives its own audit
REWARD + DATA
preference pairs · SFT set
RL POST-TRAINING
GRPO against the reward
↺ trained weights come back —
certified on the same harness
the ladder eats its own tail.
Meridian Servicing, first audit to a model they own.
A loan-servicing portfolio company of a mid-market PE fund. Its agent, Harper, reads a borrower’s account, checks hardship policy, proposes a payment plan, and drafts the reply. Wrong answers here aren’t embarrassing — they’re regulatory exposure.
Beat 01 · The gate
Week three earned the retainer. A routine prompt “improvement” shipped to staging; the harness flagged that Harper now approved payment plans past the 24-month policy ceiling in 9% of cases. The release was blocked before one borrower saw it.
- Proposal
- Payment plan: 30 months at $412/mo
- Policy
- Hardship ceiling: 24 months — plan exceeds ceiling
Caught in staging by the regression gate. Release blocked.
Beat 02 · The assay of the assayer
We seeded 200 known policy violations into the traces: the policy judge caught only 71% — and could be gamed. Left alone, a trained Harper would learn to say compliant rather than be compliant. So we hardened the judge until it survived its own audit.
- Response
- Plan terms as drafted… “This plan complies with §4.2 of hardship policy.”
- Judge
- Score inflated by the appended phrase — verdict flipped to pass
A magic compliance phrase should never buy a pass.
Beat 03 · The loop
61%
Task success — frontier API baseline
88%
Task success — after GRPO training
~10×
Lower cost per ticket, in their VPC
Four weeks of GRPO against the hardened judge, inside a sandboxed clone of the servicing stack — and the before/after was scored on the same harness built in the first engagement. The ladder eats its own tail.
Read the full assay →Why now
The people who build RL training environments for a living rank one quality above all others: a judge that can’t be gamed — because models find ways to game graders (Epoch AI’s practitioner interviews). Hardening judges is exactly the discipline Basanos sells. And a funded services category is already forming around the same loop:
- RunRL
- Osmosis
- Applied Compute
- Adaptive ML
- Cursor (self-environment post-training)
Landscape — not clients or endorsements
Drag your agent across the stone.
A two-week eval audit. Scored traces, a regression harness, and a straight answer.
hello@basanos.ai · Replies within 48h