--- schema: 1 type: video title: >- Morgan Stanley's ALPHALAB: Multi-Agent Research Across Optimization Domains — Brendan Rappazzo description: >- When coding agents got good enough at the end of 2025, Morgan Stanley's roughly thirty person research group asked what would happen if agents ran the research, not just wrote the code. The result is AlphaLab, a multi agent system they built and open sourced. You hand it a problem in plain language and it writes the code, sets up back tests and evals, configures and submits cluster jobs, and runs the statistical tests, managing its own context as it goes. They skipped every off the shelf framework and wrote their own harness so they could watch how it reasons and bake in their own standards. The shape is a strategist that proposes experiments and workers that run them, laid out as a board of cards people can read, edit, and approve before the loop optimizes against a weak eval. Rappazzo shows it finding real gains now used internally, from fine tuning a model to predicting credit bonds, and argues the lasting human job is designing the verifiable environment the agents compete in, like a private Kaggle, while they handle the middle. Speaker info: - https://x.com/brendanh0gan - https://www.linkedin.com/in/brendan-rappazzo-hogan-763734115/ - https://www.bhogan.net Timestamps: 0:00 - Introduction: a thirty person research group 1:30 - What changed when coding agents arrived 2:55 - Building AlphaLab 1.0, open sourced 4:34 - Encoding enterprise standards into the system 6:40 - Why they skipped off the shelf harnesses 8:21 - How the research loop runs 9:36 - Strategist and worker agents 10:40 - Guarding against a bad eval 13:23 - Finding real improvements 16:12 - Verifiable environments, like Kaggle 18:45 - The human job in the limit canonical: 'https://liquidgold.cc/watch/kiqubc5b5Yo' markdown: 'https://liquidgold.cc/watch/kiqubc5b5Yo.md' llms: 'https://liquidgold.cc/watch/kiqubc5b5Yo/llms.txt' parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T06:29:52.092Z' language: null children: [] videoId: kiqubc5b5Yo durationSeconds: 1207 publishedAt: '2026-07-29T17:06:26.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 12 endSeconds: 24 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=12' heldReason: rights - ordinal: 1 text: null startSeconds: 24 endSeconds: 91 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=24' heldReason: rights - ordinal: 2 text: null startSeconds: 91 endSeconds: 101 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=91' heldReason: rights - ordinal: 3 text: null startSeconds: 101 endSeconds: 177 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=101' heldReason: rights - ordinal: 4 text: null startSeconds: 101 endSeconds: 177 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=101' heldReason: rights - ordinal: 5 text: null startSeconds: 177 endSeconds: 188 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=177' heldReason: rights - ordinal: 6 text: null startSeconds: 188 endSeconds: 277 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=188' heldReason: rights - ordinal: 7 text: null startSeconds: 277 endSeconds: 290 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=277' heldReason: rights - ordinal: 8 text: null startSeconds: 290 endSeconds: 400 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=290' heldReason: rights - ordinal: 9 text: null startSeconds: 400 endSeconds: 412 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=400' heldReason: rights - ordinal: 10 text: null startSeconds: 412 endSeconds: 503 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=412' heldReason: rights - ordinal: 11 text: null startSeconds: 503 endSeconds: 514 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=503' heldReason: rights - ordinal: 12 text: null startSeconds: 514 endSeconds: 579 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=514' heldReason: rights - ordinal: 13 text: null startSeconds: 590 endSeconds: 645 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=590' heldReason: rights - ordinal: 14 text: null startSeconds: 590 endSeconds: 645 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=590' heldReason: rights - ordinal: 15 text: null startSeconds: 656 endSeconds: 806 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=656' heldReason: rights - ordinal: 16 text: null startSeconds: 656 endSeconds: 806 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=656' heldReason: rights - ordinal: 17 text: null startSeconds: 656 endSeconds: 806 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=656' heldReason: rights - ordinal: 18 text: null startSeconds: 656 endSeconds: 806 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=656' heldReason: rights - ordinal: 19 text: null startSeconds: 656 endSeconds: 806 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=656' heldReason: rights - ordinal: 20 text: null startSeconds: 806 endSeconds: 818 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=806' heldReason: rights - ordinal: 21 text: null startSeconds: 818 endSeconds: 981 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=818' heldReason: rights - ordinal: 22 text: null startSeconds: 818 endSeconds: 981 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=818' heldReason: rights - ordinal: 23 text: null startSeconds: 818 endSeconds: 981 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=818' heldReason: rights - ordinal: 24 text: null startSeconds: 991 endSeconds: 1135 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=991' heldReason: rights - ordinal: 25 text: null startSeconds: 991 endSeconds: 1135 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=991' heldReason: rights - ordinal: 26 text: null startSeconds: 991 endSeconds: 1135 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=991' heldReason: rights - ordinal: 27 text: null startSeconds: 991 endSeconds: 1135 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=991' heldReason: rights - ordinal: 28 text: null startSeconds: 1135 endSeconds: 1146 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=1135' heldReason: rights - ordinal: 29 text: null startSeconds: 1146 endSeconds: 1207 watch: 'https://liquidgold.cc/watch/kiqubc5b5Yo?t=1146' heldReason: rights --- # Morgan Stanley's ALPHALAB: Multi-Agent Research Across Optimization Domains — Brendan Rappazzo When coding agents got good enough at the end of 2025, Morgan Stanley's roughly thirty person research group asked what would happen if agents ran the research, not just wrote the code. The result is AlphaLab, a multi agent system they built and open sourced. You hand it a problem in plain language and it writes the code, sets up back tests and evals, configures and submits cluster jobs, and runs the statistical tests, managing its own context as it goes. They skipped every off the shelf framework and wrote their own harness so they could watch how it reasons and bake in their own standards. The shape is a strategist that proposes experiments and workers that run them, laid out as a board of cards people can read, edit, and approve before the loop optimizes against a weak eval. Rappazzo shows it finding real gains now used internally, from fine tuning a model to predicting credit bonds, and argues the lasting human job is designing the verifiable environment the agents compete in, like a private Kaggle, while they handle the middle. Speaker info: - https://x.com/brendanh0gan - https://www.linkedin.com/in/brendan-rappazzo-hogan-763734115/ - https://www.bhogan.net Timestamps: 0:00 - Introduction: a thirty person research group 1:30 - What changed when coding agents arrived 2:55 - Building AlphaLab 1.0, open sourced 4:34 - Encoding enterprise standards into the system 6:40 - Why they skipped off the shelf harnesses 8:21 - How the research loop runs 9:36 - Strategist and worker agents 10:40 - Guarding against a bad eval 13:23 - Finding real improvements 16:12 - Verifiable environments, like Kaggle 18:45 - The human job in the limit Canonical: https://liquidgold.cc/watch/kiqubc5b5Yo · 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.