Some results have saved runs. Some are reported from my work. Some are estimates. Here is the basis for each, including what the number does not tell you.
16 entries
Reported time-to-first-speech comparison following request parallelization, pooled HTTP connections, and end-of-turn tuning. The sampling window and percentile are unspecified.
Original resume reports a 63% reduction after moving STT, LLM, and TTS onto LiveKit's native Inference path and improving prompt caching. No matched billing window is attached to this comparison.
Hit rate reported in the original resume's voice-cost result. Its denominator and measurement window are unspecified; it is not interchangeable with a separate invoice's cached-token share.
Original resume compares the earlier Drive walk with the document-sync result. These are reported timings without an attached workload, sample size, or matched run.
Cost reduction reported alongside the document-sync timing in the original resume. The underlying billing window and workload are not specified.
Original resume outcome for the MCP agent layer and embedding-based tool retrieval. The reviewed baseline is recorded separately; no paired after-evaluation artifact is supplied for the 99% figure.
75 passing prompts divided by 229 reviewed prompts, rounded to one decimal place. This supports the baseline only, not a subsequent accuracy result.
Tool catalog size reported in the original resume, with embedding-based retrieval selecting relevant tools. This is an implementation-scale figure, not an accuracy measurement or a current registry audit.
Original resume reports organization-scoped, row-level-security-enforced sessions across a nine-organization Postgres estate, followed by query-plan fixes. The reported handler timings have no specified percentile or load.
Arithmetic means of the stored pk_wer and pk_wer2 scores for the 60 digit rows: 0.302882 and 0.052357. The same transcripts span six audio conditions. This measures a scoring correction, not new recognition performance; the number-aware scorer is not included in the retained probe.
Arithmetic means of stored dg_wer2 and pk_wer2 scores across all 240 paired rows. The pool includes 16 kHz clean audio, 8 kHz telephony, and four noise levels; it is not the separate 8 kHz summary or a production-call evaluation.
Historical report of the same seven synthesized name clips processed through 8 kHz mu-law audio, with and without directory keyterms. This is a name-subset comparison, not full-corpus WER or measured live-call routing success.
The historical report projects the seven-clip name improvement across the full corpus. The accompanying 204 ms latency is inherited from the Deepgram baseline, not a new timing measurement.
The original resume describes four verdict categories, a six-strategy fuzzy matcher, and four-level JSON repair. These are reported implementation features, not measured fact-checking accuracy.
Assumed energy constants: (1 - 0.0001 / 0.007) * 100 = 98.6%, rounded. The 0.386 kg CO2/kWh grid factor converts energy to carbon; it does not measure or validate energy consumption. No hardware energy measurement was made.
Assumed per-query energy for the smaller versus larger model: (1 - 0.0007 / 0.007) * 100 = 90%. This is the model's right-sizing scenario, not measured energy or an observed average saving.