--- schema: 1 type: video title: >- Your Coding Agent Should Do AI System Engineering — Ben Burtenshaw, Hugging Face description: >- An agent written RMSNorm kernel hit 1.88x speedups on H100s. A finetuned Qwen3 0.6B hit 35% on LiveCodeBench. Neither result required a systems engineer. Just coding agents with the right skills loaded. Ben Burtenshaw from Hugging Face walks through three levels: using Claude Code interactively to write and benchmark CUDA kernels distributed as versioned repos on the Hub, a zero-shot task where an agent finetunes a model end to end from a single prompt, and a multi agent research lab running parallel experiments overnight on Hub compute while a reporter agent pushes results to a live Trackio dashboard. The through line is skills: file based context that turns a zero shot failure into a few shot workflow. CUDA programming and ML training pipelines were deep specializations that took years. Skills compress that timeline to hours. Speaker info: - https://x.com/ben_burtenshaw - https://www.linkedin.com/in/ben-burtenshaw/ - https://github.com/burtenshaw Timestamps: 0:00 Introduction to AI Systems Engineering 1:59 Boss 1: Writing and Distributing CUDA Kernels 3:48 Efficiency in Deep Learning 6:08 Using Skills for Agentic Workflows 8:37 Benchmarking and Evaluating Skills with Upskill 9:26 Boss 2: End-to-End Fine-tuning of LLMs 10:16 Boss 3: Multi-Agent Auto Research Labs 12:09 Architecture of the Multi-Agent Research System 13:40 Implementing the Research Agent in OpenCode 15:28 Monitoring Experiments with Trackio 16:45 Final Takeaways and Conclusion --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/JomVvNDjGb8-your-coding-agent-should-do-ai-system-engineering 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/watch/JomVvNDjGb8' markdown: 'https://liquidgold.cc/watch/JomVvNDjGb8.md' llms: 'https://liquidgold.cc/watch/JomVvNDjGb8/llms.txt' parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T13:25:25.106Z' language: null children: [] videoId: JomVvNDjGb8 durationSeconds: 1105 publishedAt: '2026-05-21T13:00:06.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 26 endSeconds: 125 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=26' heldReason: rights - ordinal: 1 text: null startSeconds: 26 endSeconds: 125 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=26' heldReason: rights - ordinal: 2 text: null startSeconds: 26 endSeconds: 125 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=26' heldReason: rights - ordinal: 3 text: null startSeconds: 135 endSeconds: 234 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=135' heldReason: rights - ordinal: 4 text: null startSeconds: 135 endSeconds: 234 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=135' heldReason: rights - ordinal: 5 text: null startSeconds: 246 endSeconds: 369 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=246' heldReason: rights - ordinal: 6 text: null startSeconds: 246 endSeconds: 369 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=246' heldReason: rights - ordinal: 7 text: null startSeconds: 246 endSeconds: 369 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=246' heldReason: rights - ordinal: 8 text: null startSeconds: 380 endSeconds: 528 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=380' heldReason: rights - ordinal: 9 text: null startSeconds: 380 endSeconds: 528 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=380' heldReason: rights - ordinal: 10 text: null startSeconds: 528 endSeconds: 539 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=528' heldReason: rights - ordinal: 11 text: null startSeconds: 583 endSeconds: 620 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=583' heldReason: rights - ordinal: 12 text: null startSeconds: 631 endSeconds: 731 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=631' heldReason: rights - ordinal: 13 text: null startSeconds: 631 endSeconds: 731 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=631' heldReason: rights - ordinal: 14 text: null startSeconds: 741 endSeconds: 830 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=741' heldReason: rights - ordinal: 15 text: null startSeconds: 949 endSeconds: 1005 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=949' heldReason: rights - ordinal: 16 text: null startSeconds: 1015 endSeconds: 1105 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=1015' heldReason: rights - ordinal: 17 text: null startSeconds: 1015 endSeconds: 1105 watch: 'https://liquidgold.cc/watch/JomVvNDjGb8?t=1015' heldReason: rights --- # Your Coding Agent Should Do AI System Engineering — Ben Burtenshaw, Hugging Face An agent written RMSNorm kernel hit 1.88x speedups on H100s. A finetuned Qwen3 0.6B hit 35% on LiveCodeBench. Neither result required a systems engineer. Just coding agents with the right skills loaded. Ben Burtenshaw from Hugging Face walks through three levels: using Claude Code interactively to write and benchmark CUDA kernels distributed as versioned repos on the Hub, a zero-shot task where an agent finetunes a model end to end from a single prompt, and a multi agent research lab running parallel experiments overnight on Hub compute while a reporter agent pushes results to a live Trackio dashboard. The through line is skills: file based context that turns a zero shot failure into a few shot workflow. CUDA programming and ML training pipelines were deep specializations that took years. Skills compress that timeline to hours. Speaker info: - https://x.com/ben_burtenshaw - https://www.linkedin.com/in/ben-burtenshaw/ - https://github.com/burtenshaw Timestamps: 0:00 Introduction to AI Systems Engineering 1:59 Boss 1: Writing and Distributing CUDA Kernels 3:48 Efficiency in Deep Learning 6:08 Using Skills for Agentic Workflows 8:37 Benchmarking and Evaluating Skills with Upskill 9:26 Boss 2: End-to-End Fine-tuning of LLMs 10:16 Boss 3: Multi-Agent Auto Research Labs 12:09 Architecture of the Multi-Agent Research System 13:40 Implementing the Research Agent in OpenCode 15:28 Monitoring Experiments with Trackio 16:45 Final Takeaways and Conclusion --- Watch along with the corrected transcript, summary, timestamps and resources on this talk's official AIE page: https://ai.engineer/talks/JomVvNDjGb8-your-coding-agent-should-do-ai-system-engineering 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/watch/JomVvNDjGb8 · 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.