--- schema: 1 type: video title: 'Frontier results, on device - RL Nabors, Arize' description: >- Most of use reach for a frontier model by default and pay for it on every call, in latency, in energy, in cash, and in everything that leaves their stack. For most of those calls, a small local model would do the job. RL Nabors, former Meta/React core team member and AWS alum, covers the vocabulary you need to reason about model performance (capability evals, golden datasets, LLM-as-judge) and walks through real cases: a local agentic harness replacing a frontier call, an in-browser moderation classifier defended with production-trace evals, and a generative summarization feature where the rubric turns out to be harder than the model. You'll leave with a framework for deciding when to choose large and off-prem or small and local models, and how to measure your way to the answer instead of guessing. You will learn: - The vocabulary to reason about model performance (capability evals, golden datasets, LLM-as-judge). - A framework for deciding when a small or local model can replace a frontier one and when it can't. - A repeatable process for building capability evals from your own production traces, not someone else's benchmark. - Working examples of using eval results to iterate on prompts and ship with confidence instead of vibes. Speakers: - RL Nabors (Arize): RL Nabors builds developer tools and the communities that make them stick. Previously React and MDN, currently developer experience at Arize, perpetually building Mima. X/Twitter: https://x.com/rachelnabors LinkedIn: https://linkedin.com/in/nearestnabors GitHub: https://linkedin.com/in/nearestnabors --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/fWXJM-J0ZB8-frontier-results-on-device-rl-nabors-arize Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/frontier-results-on-device-rl-nabors-arize-6efd941dff16 markdown: >- https://liquidgold.cc/channel/aidotengineer/video/frontier-results-on-device-rl-nabors-arize-6efd941dff16.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T08:55:41.608Z' language: null children: [] videoId: fWXJM-J0ZB8 durationSeconds: 1852 publishedAt: '2026-06-29T03:30:03.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 11 endSeconds: 126 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=11' heldReason: rights - ordinal: 1 text: null startSeconds: 11 endSeconds: 126 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=11' heldReason: rights - ordinal: 2 text: null startSeconds: 11 endSeconds: 126 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=11' heldReason: rights - ordinal: 3 text: null startSeconds: 137 endSeconds: 223 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=137' heldReason: rights - ordinal: 4 text: null startSeconds: 137 endSeconds: 223 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=137' heldReason: rights - ordinal: 5 text: null startSeconds: 235 endSeconds: 335 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=235' heldReason: rights - ordinal: 6 text: null startSeconds: 235 endSeconds: 335 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=235' heldReason: rights - ordinal: 7 text: null startSeconds: 345 endSeconds: 467 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=345' heldReason: rights - ordinal: 8 text: null startSeconds: 345 endSeconds: 467 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=345' heldReason: rights - ordinal: 9 text: null startSeconds: 477 endSeconds: 589 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=477' heldReason: rights - ordinal: 10 text: null startSeconds: 601 endSeconds: 714 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=601' heldReason: rights - ordinal: 11 text: null startSeconds: 714 endSeconds: 726 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=714' heldReason: rights - ordinal: 12 text: null startSeconds: 726 endSeconds: 808 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=726' heldReason: rights - ordinal: 13 text: null startSeconds: 819 endSeconds: 919 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=819' heldReason: rights - ordinal: 14 text: null startSeconds: 819 endSeconds: 919 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=819' heldReason: rights - ordinal: 15 text: null startSeconds: 929 endSeconds: 1041 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=929' heldReason: rights - ordinal: 16 text: null startSeconds: 929 endSeconds: 1041 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=929' heldReason: rights - ordinal: 17 text: null startSeconds: 1053 endSeconds: 1164 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1053' heldReason: rights - ordinal: 18 text: null startSeconds: 1175 endSeconds: 1296 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1175' heldReason: rights - ordinal: 19 text: null startSeconds: 1307 endSeconds: 1392 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1307' heldReason: rights - ordinal: 20 text: null startSeconds: 1403 endSeconds: 1516 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1403' heldReason: rights - ordinal: 21 text: null startSeconds: 1527 endSeconds: 1617 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1527' heldReason: rights - ordinal: 22 text: null startSeconds: 1617 endSeconds: 1629 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1617' heldReason: rights - ordinal: 23 text: null startSeconds: 1629 endSeconds: 1709 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1629' heldReason: rights - ordinal: 24 text: null startSeconds: 1720 endSeconds: 1852 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1720' heldReason: rights - ordinal: 25 text: null startSeconds: 1720 endSeconds: 1852 watch: 'https://liquidgold.cc/watch/fWXJM-J0ZB8?t=1720' heldReason: rights --- # Frontier results, on device - RL Nabors, Arize Most of use reach for a frontier model by default and pay for it on every call, in latency, in energy, in cash, and in everything that leaves their stack. For most of those calls, a small local model would do the job. RL Nabors, former Meta/React core team member and AWS alum, covers the vocabulary you need to reason about model performance (capability evals, golden datasets, LLM-as-judge) and walks through real cases: a local agentic harness replacing a frontier call, an in-browser moderation classifier defended with production-trace evals, and a generative summarization feature where the rubric turns out to be harder than the model. You'll leave with a framework for deciding when to choose large and off-prem or small and local models, and how to measure your way to the answer instead of guessing. You will learn: - The vocabulary to reason about model performance (capability evals, golden datasets, LLM-as-judge). - A framework for deciding when a small or local model can replace a frontier one and when it can't. - A repeatable process for building capability evals from your own production traces, not someone else's benchmark. - Working examples of using eval results to iterate on prompts and ship with confidence instead of vibes. Speakers: - RL Nabors (Arize): RL Nabors builds developer tools and the communities that make them stick. Previously React and MDN, currently developer experience at Arize, perpetually building Mima. X/Twitter: https://x.com/rachelnabors LinkedIn: https://linkedin.com/in/nearestnabors GitHub: https://linkedin.com/in/nearestnabors --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/fWXJM-J0ZB8-frontier-results-on-device-rl-nabors-arize Join our upcoming AI Engineer conferences (see https://ai.engineer landing page): - AIEi Shanghai, November 5–6 — the first for meeting top Chinese labs and engineers - AIE CODE in SF, November 10–12 — the top AI Coding conference returns - AIEi Sydney, December 7–8 — the first AIEi run alongside NeurIPS 2026 in Sydney --- Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/frontier-results-on-device-rl-nabors-arize-6efd941dff16 · Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks > Missing: no stored document chunks are available.