--- schema: 1 type: video title: >- Memory-Aware Agent Harnesses for Anti-Money Laundering — Nacho Martinez, Oracle description: "The model is the part you can't change. The harness around it is where you can make real gains. Nacho Martinez walks through the full agent stack (agent loop, short- and long-term memory, context engineering, the toolbox and skillbox patterns, skill distillation and continual learning) and applies it to an autonomous agent that flags suspicious bank transactions around the clock. \U0001F39F️ Join us at the next one: AIE CODE Summit → https://aie.codes Speaker Nacho Martinez — LinkedIn: https://www.linkedin.com/in/jasperan/ GitHub: https://github.com/jasperan Oracle — https://www.oracle.com Workshop: https://workshopwaitingroom.com Timestamps 0:00 Workshop setup 2:07 Intro: Nacho Martinez 2:57 Agenda 4:57 The use case: an AML agent for a mid-sized bank 5:57 The agent stack 8:31 Agentic systems = automation + autonomy 9:16 What is an AI agent? 11:06 Why the harness matters more than the model 13:21 Files vs. databases 15:56 Agent memory: short-term, episodic, procedural, semantic 19:01 Context summarization vs. compaction 19:46 The agent loop 20:51 Toolbox & skillbox patterns 22:30 Skill distillation 24:15 How a frozen model gets better: continual learning 26:10 Instruction hierarchy 27:05 Workshop UI walkthrough 33:35 Q&A: Episodic memory & conflicts 34:35 Q&A: Skills, caching & context budgets 40:49 Q&A: Keep your memory layer simple 43:29 Q&A: Pre-built SQL vs. dynamic queries 46:14 Hybrid search (semantic + keyword) 47:39 Build small custom harnesses" canonical: 'https://liquidgold.cc/watch/KtBivPzmOM8' markdown: 'https://liquidgold.cc/watch/KtBivPzmOM8.md' llms: 'https://liquidgold.cc/watch/KtBivPzmOM8/llms.txt' parent: 'https://liquidgold.cc/channel/aidotengineer.md' channel: aidotengineer updated: '2026-10-23T22:23:41.837Z' language: null children: [] videoId: KtBivPzmOM8 durationSeconds: 3109 publishedAt: '2026-10-17T07:00:25.000Z' searchChunks: - ordinal: 0 text: null startSeconds: 197 endSeconds: 300 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=197' heldReason: rights - ordinal: 1 text: null startSeconds: 300 endSeconds: 310 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=300' heldReason: rights - ordinal: 2 text: null startSeconds: 379 endSeconds: 513 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=379' heldReason: rights - ordinal: 3 text: null startSeconds: 513 endSeconds: 526 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=513' heldReason: rights - ordinal: 4 text: null startSeconds: 572 endSeconds: 675 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=572' heldReason: rights - ordinal: 5 text: null startSeconds: 686 endSeconds: 810 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=686' heldReason: rights - ordinal: 6 text: null startSeconds: 821 endSeconds: 958 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=821' heldReason: rights - ordinal: 7 text: null startSeconds: 821 endSeconds: 958 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=821' heldReason: rights - ordinal: 8 text: null startSeconds: 968 endSeconds: 1149 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=968' heldReason: rights - ordinal: 9 text: null startSeconds: 968 endSeconds: 1149 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=968' heldReason: rights - ordinal: 10 text: null startSeconds: 1149 endSeconds: 1162 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1149' heldReason: rights - ordinal: 11 text: null startSeconds: 1186 endSeconds: 1197 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1186' heldReason: rights - ordinal: 12 text: null startSeconds: 1262 endSeconds: 1350 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1262' heldReason: rights - ordinal: 13 text: null startSeconds: 1350 endSeconds: 1360 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1350' heldReason: rights - ordinal: 14 text: null startSeconds: 1459 endSeconds: 1470 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1459' heldReason: rights - ordinal: 15 text: null startSeconds: 1470 endSeconds: 1572 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1470' heldReason: rights - ordinal: 16 text: null startSeconds: 1584 endSeconds: 1628 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1584' heldReason: rights - ordinal: 17 text: null startSeconds: 1770 endSeconds: 1872 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=1770' heldReason: rights - ordinal: 18 text: null startSeconds: 2079 endSeconds: 2091 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2079' heldReason: rights - ordinal: 19 text: null startSeconds: 2199 endSeconds: 2322 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2199' heldReason: rights - ordinal: 20 text: null startSeconds: 2333 endSeconds: 2457 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2333' heldReason: rights - ordinal: 21 text: null startSeconds: 2457 endSeconds: 2469 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2457' heldReason: rights - ordinal: 22 text: null startSeconds: 2618 endSeconds: 2628 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2618' heldReason: rights - ordinal: 23 text: null startSeconds: 2774 endSeconds: 2785 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2774' heldReason: rights - ordinal: 24 text: null startSeconds: 2862 endSeconds: 2873 watch: 'https://liquidgold.cc/watch/KtBivPzmOM8?t=2862' heldReason: rights --- # Memory-Aware Agent Harnesses for Anti-Money Laundering — Nacho Martinez, Oracle The model is the part you can't change. The harness around it is where you can make real gains. Nacho Martinez walks through the full agent stack (agent loop, short- and long-term memory, context engineering, the toolbox and skillbox patterns, skill distillation and continual learning) and applies it to an autonomous agent that flags suspicious bank transactions around the clock. 🎟️ Join us at the next one: AIE CODE Summit → https://aie.codes Speaker Nacho Martinez — LinkedIn: https://www.linkedin.com/in/jasperan/ GitHub: https://github.com/jasperan Oracle — https://www.oracle.com Workshop: https://workshopwaitingroom.com Timestamps 0:00 Workshop setup 2:07 Intro: Nacho Martinez 2:57 Agenda 4:57 The use case: an AML agent for a mid-sized bank 5:57 The agent stack 8:31 Agentic systems = automation + autonomy 9:16 What is an AI agent? 11:06 Why the harness matters more than the model 13:21 Files vs. databases 15:56 Agent memory: short-term, episodic, procedural, semantic 19:01 Context summarization vs. compaction 19:46 The agent loop 20:51 Toolbox & skillbox patterns 22:30 Skill distillation 24:15 How a frozen model gets better: continual learning 26:10 Instruction hierarchy 27:05 Workshop UI walkthrough 33:35 Q\&A: Episodic memory & conflicts 34:35 Q\&A: Skills, caching & context budgets 40:49 Q\&A: Keep your memory layer simple 43:29 Q\&A: Pre-built SQL vs. dynamic queries 46:14 Hybrid search (semantic + keyword) 47:39 Build small custom harnesses Canonical: https://liquidgold.cc/watch/KtBivPzmOM8 · 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 (0:12 to 3:06) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 1: Document excerpt (3:06 to 5:10) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 2: Document excerpt (5:10 to 6:19) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 3: Document excerpt (6:19 to 8:33) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 4: Document excerpt (8:33 to 9:19) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 5: Document excerpt (9:19 to 11:15) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 6: Document excerpt (11:15 to 13:30) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 7: Document excerpt (13:30 to 15:58) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 8: Document excerpt (15:58 to 19:09) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 9: Document excerpt (19:09 to 19:46) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 10: Document excerpt (19:46 to 20:52) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 11: Document excerpt (20:52 to 22:30) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 12: Document excerpt (22:30 to 24:19) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 13: Document excerpt (24:19 to 26:12) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 14: Document excerpt (26:12 to 27:08) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 15: Document excerpt (27:08 to 29:19) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 16: Document excerpt (29:19 to 31:24) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 17: Document excerpt (31:24 to 33:44) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 18: Document excerpt (33:44 to 34:39) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 19: Document excerpt (34:39 to 36:29) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 20: Document excerpt (36:29 to 38:42) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 21: Document excerpt (38:42 to 40:57) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 22: Document excerpt (40:57 to 43:38) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 23: Document excerpt (43:38 to 46:14) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 24: Document excerpt (46:14 to 47:42) > Held: transcript rights do not permit this chunk's text. --- #### Chunk 25: Document excerpt (47:42 to 51:49) > Held: transcript rights do not permit this chunk's text.