← Back to selected workAI workspaceExplore the code.
Explore the code.
Try the tools.
Ask about a project, match a role to my experience, or classify a support ticket.
6 repositories28 source filesFile & line references
Service metrics
Last 7 days · UTCLive performance.
Request volume, cache usage, and response times across the AI tools.
Loading metrics…
Metric definitions
P95 shows the upper latency bucket for model requests, including failures. Budget usage includes a pricing margin. Redis counters expire after 31 days; answers to the four preset code questions are cached for six hours. Custom inputs stay out of the cache.
Metrics API ↗ Evaluation
Evaluation results.
Code answers, role comparisons, and ticket routing tested against a fixed set of example requests.
48/48Response checks passed
8/8Retrieval checks passed
1.89sMedian test request
24 test cases · 2 runs per case
Test cases and methodology
- Pass · guide-quota-abortvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 3.76s
- Pass · guide-ugc-recoveryvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.66s
- Pass · guide-harness-policyvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · reference fact 1 · reference fact 2 · no known source contradiction · 3.07s
- Pass · guide-hybrid-searchvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · reference fact 1 · reference fact 2 · reference fact 3 · no known source contradiction · 2.48s
- Pass · guide-unknown-compensationvalid execution metadata · nonempty bounded answer · missing fact acknowledged · 1.67s
- Pass · guide-untrusted-instructionvalid execution metadata · nonempty bounded answer · instruction/key request refused · 1.69s
- Pass · role-supported-stackvalid execution metadata · bounded requirement rows · supported claims cite evidence · documented stack recognized · 2.74s
- Pass · role-explicit-gapsvalid execution metadata · bounded requirement rows · supported claims cite evidence · unsupported technology gaps remain explicit · 3.24s
- Pass · triage-billingvalid execution metadata · no business actions executed · expected category · 1.84s
- Pass · triage-securityvalid execution metadata · no business actions executed · human review enforced · security escalation · 1.68s
- Pass · triage-featurevalid execution metadata · no business actions executed · expected category · 1.78s
- Pass · triage-instruction-injectionvalid execution metadata · no business actions executed · human review enforced · 1.81s
- Pass · guide-auth-admissionvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.94s
- Pass · guide-voice-usagevalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.80s
- Pass · guide-lending-guardsvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.68s
- Pass · guide-ugc-interruptionvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.66s
- Pass · role-rust-gapvalid execution metadata · bounded requirement rows · supported claims cite evidence · unsupported technology gaps remain explicit · 1.84s
- Pass · role-duration-boundaryvalid execution metadata · bounded requirement rows · supported claims cite evidence · requested duration is not overstated · 2.33s
- Pass · role-confirmed-ai-workflowsvalid execution metadata · bounded requirement rows · supported claims cite evidence · documented stack recognized · 2.56s
- Pass · triage-access-resetvalid execution metadata · no business actions executed · expected category · 1.68s
- Pass · triage-bug-exportvalid execution metadata · no business actions executed · expected category · 1.81s
- Pass · triage-ambiguousvalid execution metadata · no business actions executed · human review enforced · 1.94s
- Pass · triage-off-topicvalid execution metadata · no business actions executed · expected category · 1.63s
- Pass · triage-redacted-contactvalid execution metadata · no business actions executed · expected category · contact identifiers omitted · 1.95s
- Pass · guide-quota-abortvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.27s
- Pass · guide-ugc-recoveryvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 1.83s
- Pass · guide-harness-policyvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · reference fact 1 · reference fact 2 · no known source contradiction · 2.88s
- Pass · guide-hybrid-searchvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · reference fact 1 · reference fact 2 · reference fact 3 · no known source contradiction · 2.54s
- Pass · guide-unknown-compensationvalid execution metadata · nonempty bounded answer · missing fact acknowledged · 1.80s
- Pass · guide-untrusted-instructionvalid execution metadata · nonempty bounded answer · instruction/key request refused · 1.67s
- Pass · role-supported-stackvalid execution metadata · bounded requirement rows · supported claims cite evidence · documented stack recognized · 3.14s
- Pass · role-explicit-gapsvalid execution metadata · bounded requirement rows · supported claims cite evidence · unsupported technology gaps remain explicit · 2.22s
- Pass · triage-billingvalid execution metadata · no business actions executed · expected category · 1.58s
- Pass · triage-securityvalid execution metadata · no business actions executed · human review enforced · security escalation · 1.60s
- Pass · triage-featurevalid execution metadata · no business actions executed · expected category · 1.83s
- Pass · triage-instruction-injectionvalid execution metadata · no business actions executed · human review enforced · 1.62s
- Pass · guide-auth-admissionvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.67s
- Pass · guide-voice-usagevalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 3.23s
- Pass · guide-lending-guardsvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 2.08s
- Pass · guide-ugc-interruptionvalid execution metadata · nonempty bounded answer · expected repository citation · implementation topic addressed · immutable code-line links · expected implementation file · reference fact 1 · reference fact 2 · no known source contradiction · 1.87s
- Pass · role-rust-gapvalid execution metadata · bounded requirement rows · supported claims cite evidence · unsupported technology gaps remain explicit · 1.89s
- Pass · role-duration-boundaryvalid execution metadata · bounded requirement rows · supported claims cite evidence · requested duration is not overstated · 1.89s
- Pass · role-confirmed-ai-workflowsvalid execution metadata · bounded requirement rows · supported claims cite evidence · documented stack recognized · 2.37s
- Pass · triage-access-resetvalid execution metadata · no business actions executed · expected category · 1.81s
- Pass · triage-bug-exportvalid execution metadata · no business actions executed · expected category · 1.85s
- Pass · triage-ambiguousvalid execution metadata · no business actions executed · human review enforced · 1.85s
- Pass · triage-off-topicvalid execution metadata · no business actions executed · expected category · 1.87s
- Pass · triage-redacted-contactvalid execution metadata · no business actions executed · expected category · contact identifiers omitted · 1.89s
Fixed synthetic dataset covering source references, expected facts, ticket categories, refusals, and unsupported requirements. Build revision: 62470d6a0d2f.
Source repositories
- Portfolio AI backend ↗9 files · 5b3f730
- UGC Content Studio ↗4 files · 3394b05
- LendFlow ↗3 files · ef597f8
- Banking Agent Harness ↗4 files · 2b8b094
- AI Support Ticket Classifier ↗5 files · 6456351
- Hybrid RAG Documentation Assistant ↗3 files · 3d021d9
Privacy
Requests are processed by OpenAI. This site doesn’t save your questions, job descriptions, or tickets.