--- schema: 1 type: video title: 'Autonomous Agents for Scientific Tasks - Sina Shahandeh, Radicait' description: >- There has been much work on Autoresearch where the objectives are coding puzzles, toy optimization problems, or static supervised-learning ML tasks. However, for an autonomous agent to assist with a scientific discovery task, the problems must come from real measurement data of the world and they are highly open-ended, requiring a scientific method in the solution loop. In this talk, we show scenarios where the agent has to search over methods, priors, data preprocessing, model classes, and hyperparameters while learning from intermediate failures. Often, a step change in agent performance comes from forming an appropriate scientific hypothesis about how the physical system behaves, implementing that hypothesis correctly in a mathematical model, and executing it on the existing real data. We show how an ontology-based memory system is used in the harness to assist with the hypothesis generation that is key to the agent's success. All demonstrations come from real scientific problems solved in industrial and applied research settings. Speakers: - Sina Shahandeh (Radicait): Sina is a cofounder/CTO at Radicait. His background is in scientific computing (PhD) and has lead data/AI function in 3 scale ups (with $770M exit) as well as founding 3 companies. X/Twitter: https://x.com/SinaShahandeh LinkedIn: https://www.linkedin.com/in/sinashahandeh/ canonical: >- https://liquidgold.cc/channel/aidotengineer/explore/video/autonomous-agents-for-scientific-tasks-sina-shahand-b750009bb616 markdown: >- https://liquidgold.cc/channel/aidotengineer/video/autonomous-agents-for-scientific-tasks-sina-shahand-b750009bb616.md llms: null parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-24T07:24:31.141Z' language: null children: [] videoId: XLEYtv3cMlw durationSeconds: 1163 publishedAt: '2026-07-18T23:30:06.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 12 endSeconds: 101 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=12' heldReason: rights - ordinal: 1 text: null startSeconds: 12 endSeconds: 101 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=12' heldReason: rights - ordinal: 2 text: null startSeconds: 112 endSeconds: 223 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=112' heldReason: rights - ordinal: 3 text: null startSeconds: 112 endSeconds: 223 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=112' heldReason: rights - ordinal: 4 text: null startSeconds: 234 endSeconds: 315 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=234' heldReason: rights - ordinal: 5 text: null startSeconds: 326 endSeconds: 428 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=326' heldReason: rights - ordinal: 6 text: null startSeconds: 326 endSeconds: 428 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=326' heldReason: rights - ordinal: 7 text: null startSeconds: 439 endSeconds: 559 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=439' heldReason: rights - ordinal: 8 text: null startSeconds: 439 endSeconds: 559 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=439' heldReason: rights - ordinal: 9 text: null startSeconds: 570 endSeconds: 664 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=570' heldReason: rights - ordinal: 10 text: null startSeconds: 570 endSeconds: 664 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=570' heldReason: rights - ordinal: 11 text: null startSeconds: 676 endSeconds: 794 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=676' heldReason: rights - ordinal: 12 text: null startSeconds: 805 endSeconds: 885 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=805' heldReason: rights - ordinal: 13 text: null startSeconds: 805 endSeconds: 885 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=805' heldReason: rights - ordinal: 14 text: null startSeconds: 896 endSeconds: 977 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=896' heldReason: rights - ordinal: 15 text: null startSeconds: 896 endSeconds: 977 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=896' heldReason: rights - ordinal: 16 text: null startSeconds: 896 endSeconds: 977 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=896' heldReason: rights - ordinal: 17 text: null startSeconds: 988 endSeconds: 1095 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=988' heldReason: rights - ordinal: 18 text: null startSeconds: 988 endSeconds: 1095 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=988' heldReason: rights - ordinal: 19 text: null startSeconds: 1107 endSeconds: 1163 watch: 'https://liquidgold.cc/watch/XLEYtv3cMlw?t=1107' heldReason: rights --- # Autonomous Agents for Scientific Tasks - Sina Shahandeh, Radicait There has been much work on Autoresearch where the objectives are coding puzzles, toy optimization problems, or static supervised-learning ML tasks. However, for an autonomous agent to assist with a scientific discovery task, the problems must come from real measurement data of the world and they are highly open-ended, requiring a scientific method in the solution loop. In this talk, we show scenarios where the agent has to search over methods, priors, data preprocessing, model classes, and hyperparameters while learning from intermediate failures. Often, a step change in agent performance comes from forming an appropriate scientific hypothesis about how the physical system behaves, implementing that hypothesis correctly in a mathematical model, and executing it on the existing real data. We show how an ontology-based memory system is used in the harness to assist with the hypothesis generation that is key to the agent's success. All demonstrations come from real scientific problems solved in industrial and applied research settings. Speakers: - Sina Shahandeh (Radicait): Sina is a cofounder/CTO at Radicait. His background is in scientific computing (PhD) and has lead data/AI function in 3 scale ups (with $770M exit) as well as founding 3 companies. X/Twitter: https://x.com/SinaShahandeh LinkedIn: https://www.linkedin.com/in/sinashahandeh/ Canonical: https://liquidgold.cc/channel/aidotengineer/explore/video/autonomous-agents-for-scientific-tasks-sina-shahand-b750009bb616 ยท Channel: [AI Engineer](https://liquidgold.cc/channel/aidotengineer.md) > Search chunk text is held until public-source provenance is verified. ## Document chunks --- #### Chunk 0: Document excerpt (0:00 to 1:41) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 1: Document excerpt (1:41 to 3:43) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 2: Document excerpt (3:43 to 5:15) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 3: Document excerpt (5:15 to 7:08) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 4: Document excerpt (7:08 to 9:19) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 5: Document excerpt (9:19 to 11:04) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 6: Document excerpt (11:04 to 13:14) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 7: Document excerpt (13:14 to 14:45) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 8: Document excerpt (14:45 to 16:17) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 9: Document excerpt (16:17 to 18:15) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 10: Document excerpt (18:15 to 19:23) > Held: transcript rights do not permit this chunk's text.