{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FUBBGDNQL4PGH2VG527J2L6QGF","short_pith_number":"pith:FUBBGDNQ","schema_version":"1.0","canonical_sha256":"2d02130db05f1e63eaa6eebe9d2fd03160c7ff696d2f08060495cfcd63d27f80","source":{"kind":"arxiv","id":"2507.16835","version":2},"attestation_state":"computed","paper":{"title":"Evaluating Speech-to-Text x LLM x Text-to-Speech Combinations for AI Interview Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"eess.AS","authors_text":"Ali Ansari, Amirhossein Afsharrad, Aruj Mahajan, Nima Yazdani, Rumi Allbert, Seyed Shahabeddin Mousavi","submitted_at":"2025-07-15T22:30:55Z","abstract_excerpt":"Voice-based conversational AI systems increasingly rely on cascaded architectures that combine speech-to-text (STT), large language models (LLMs), and text-to-speech (TTS) components. We present a large-scale empirical comparison of STT x LLM x TTS stacks using data sampled from over 300,000 AI-conducted job interviews. We used an LLM-as-a-Judge automated evaluation framework to assess conversational quality, technical accuracy, and skill assessment capabilities. Our analysis of five production configurations reveals that a stack combining Google's STT, GPT-4.1, and Cartesia's TTS outperforms "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2507.16835","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"eess.AS","submitted_at":"2025-07-15T22:30:55Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"2a74c6fbe7b1095c421c599b9074586be640e2622664f54772810a3865e01983","abstract_canon_sha256":"f68d456e21a53532de9a831d70ca2a4e5047d2d797b6bedebcc4457555bf6bb0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:57:20.735736Z","signature_b64":"2DXFpVpa895E1thGum/7hy/bjykRru0Yj55Dd1XxqKhJCFPpN+7JEg8WnmdiHJHfGz6sV0a/8e2fhPDf0VYIAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d02130db05f1e63eaa6eebe9d2fd03160c7ff696d2f08060495cfcd63d27f80","last_reissued_at":"2026-07-05T11:57:20.735222Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:57:20.735222Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating Speech-to-Text x LLM x Text-to-Speech Combinations for AI Interview Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"eess.AS","authors_text":"Ali Ansari, Amirhossein Afsharrad, Aruj Mahajan, Nima Yazdani, Rumi Allbert, Seyed Shahabeddin Mousavi","submitted_at":"2025-07-15T22:30:55Z","abstract_excerpt":"Voice-based conversational AI systems increasingly rely on cascaded architectures that combine speech-to-text (STT), large language models (LLMs), and text-to-speech (TTS) components. We present a large-scale empirical comparison of STT x LLM x TTS stacks using data sampled from over 300,000 AI-conducted job interviews. We used an LLM-as-a-Judge automated evaluation framework to assess conversational quality, technical accuracy, and skill assessment capabilities. Our analysis of five production configurations reveals that a stack combining Google's STT, GPT-4.1, and Cartesia's TTS outperforms "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.16835","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2507.16835/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2507.16835","created_at":"2026-07-05T11:57:20.735298+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.16835v2","created_at":"2026-07-05T11:57:20.735298+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.16835","created_at":"2026-07-05T11:57:20.735298+00:00"},{"alias_kind":"pith_short_12","alias_value":"FUBBGDNQL4PG","created_at":"2026-07-05T11:57:20.735298+00:00"},{"alias_kind":"pith_short_16","alias_value":"FUBBGDNQL4PGH2VG","created_at":"2026-07-05T11:57:20.735298+00:00"},{"alias_kind":"pith_short_8","alias_value":"FUBBGDNQ","created_at":"2026-07-05T11:57:20.735298+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.04556","citing_title":"Benchmarking LLMs on the Massive Sound Embedding Benchmark (MSEB)","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF","json":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF.json","graph_json":"https://pith.science/api/pith-number/FUBBGDNQL4PGH2VG527J2L6QGF/graph.json","events_json":"https://pith.science/api/pith-number/FUBBGDNQL4PGH2VG527J2L6QGF/events.json","paper":"https://pith.science/paper/FUBBGDNQ"},"agent_actions":{"view_html":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF","download_json":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF.json","view_paper":"https://pith.science/paper/FUBBGDNQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.16835&json=true","fetch_graph":"https://pith.science/api/pith-number/FUBBGDNQL4PGH2VG527J2L6QGF/graph.json","fetch_events":"https://pith.science/api/pith-number/FUBBGDNQL4PGH2VG527J2L6QGF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF/action/storage_attestation","attest_author":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF/action/author_attestation","sign_citation":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF/action/citation_signature","submit_replication":"https://pith.science/pith/FUBBGDNQL4PGH2VG527J2L6QGF/action/replication_record"}},"created_at":"2026-07-05T11:57:20.735298+00:00","updated_at":"2026-07-05T11:57:20.735298+00:00"}