{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NLRRMPO23QNLGG7TN7LMYFA7FE","short_pith_number":"pith:NLRRMPO2","schema_version":"1.0","canonical_sha256":"6ae3163ddadc1ab31bf36fd6cc141f29315c418262543a824bd670e870766cc9","source":{"kind":"arxiv","id":"2607.04395","version":1},"attestation_state":"computed","paper":{"title":"NKI-Agent: Domain-Specific Fine-Tuning and Agentic Tool Use for Neuron Kernel Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hao Zhou, Jun Huan, Junjie Tang, Lin Wang, Yuhao Zhang","submitted_at":"2026-07-05T16:37:41Z","abstract_excerpt":"Recent agentic approaches to LLM-based kernel generation have achieved impressive results on CUDA. For emerging AI accelerators such as AWS Trainium and Inferentia, automated kernel generation and optimization remain largely unaddressed. Writing kernels for these chips via the Neuron Kernel Interface (NKI) is particularly challenging: developers must navigate a multi-engine architecture, tile-based programming, and explicit data movement across multi-level memory hierarchy. Moreover, no publicly-available training data, benchmarks, or tool-augmented agents exist for this domain. We introduce N"},"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":"2607.04395","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-07-05T16:37:41Z","cross_cats_sorted":[],"title_canon_sha256":"f573326786dd3692d89f329ef3d52d2683a56aaba40053b55f7bd8b356c87ef7","abstract_canon_sha256":"d6853656eec4075f62783fca30134f0117602aabfc6d9aa3f99a61f16659f788"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:19:16.439716Z","signature_b64":"axfvJNLBQKX2URADLUM4NBvohHx5EQGss1A3eZF+gOBb9bRqLCmyli+3djQR20XK6973eyTE6xWhcSmGhy6pAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ae3163ddadc1ab31bf36fd6cc141f29315c418262543a824bd670e870766cc9","last_reissued_at":"2026-07-07T02:19:16.439019Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:19:16.439019Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NKI-Agent: Domain-Specific Fine-Tuning and Agentic Tool Use for Neuron Kernel Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hao Zhou, Jun Huan, Junjie Tang, Lin Wang, Yuhao Zhang","submitted_at":"2026-07-05T16:37:41Z","abstract_excerpt":"Recent agentic approaches to LLM-based kernel generation have achieved impressive results on CUDA. For emerging AI accelerators such as AWS Trainium and Inferentia, automated kernel generation and optimization remain largely unaddressed. Writing kernels for these chips via the Neuron Kernel Interface (NKI) is particularly challenging: developers must navigate a multi-engine architecture, tile-based programming, and explicit data movement across multi-level memory hierarchy. Moreover, no publicly-available training data, benchmarks, or tool-augmented agents exist for this domain. We introduce N"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.04395","kind":"arxiv","version":1},"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/2607.04395/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":"2607.04395","created_at":"2026-07-07T02:19:16.439108+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.04395v1","created_at":"2026-07-07T02:19:16.439108+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.04395","created_at":"2026-07-07T02:19:16.439108+00:00"},{"alias_kind":"pith_short_12","alias_value":"NLRRMPO23QNL","created_at":"2026-07-07T02:19:16.439108+00:00"},{"alias_kind":"pith_short_16","alias_value":"NLRRMPO23QNLGG7T","created_at":"2026-07-07T02:19:16.439108+00:00"},{"alias_kind":"pith_short_8","alias_value":"NLRRMPO2","created_at":"2026-07-07T02:19:16.439108+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE","json":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE.json","graph_json":"https://pith.science/api/pith-number/NLRRMPO23QNLGG7TN7LMYFA7FE/graph.json","events_json":"https://pith.science/api/pith-number/NLRRMPO23QNLGG7TN7LMYFA7FE/events.json","paper":"https://pith.science/paper/NLRRMPO2"},"agent_actions":{"view_html":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE","download_json":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE.json","view_paper":"https://pith.science/paper/NLRRMPO2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.04395&json=true","fetch_graph":"https://pith.science/api/pith-number/NLRRMPO23QNLGG7TN7LMYFA7FE/graph.json","fetch_events":"https://pith.science/api/pith-number/NLRRMPO23QNLGG7TN7LMYFA7FE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE/action/storage_attestation","attest_author":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE/action/author_attestation","sign_citation":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE/action/citation_signature","submit_replication":"https://pith.science/pith/NLRRMPO23QNLGG7TN7LMYFA7FE/action/replication_record"}},"created_at":"2026-07-07T02:19:16.439108+00:00","updated_at":"2026-07-07T02:19:16.439108+00:00"}