Uploaded July 2026 | Updated September 2026, 1 week ago

---

Materials : docs.google.com/document/d/1wMPQL2NJTzT70GLBVYr3hKrObCmYrwhwvTgoEb0PLWk/edit?usp=sharing

Here are the detailed timestamps for the session:

(0:00:00 - 0:19:45) Introduction and Mentors' Overview
(0:19:45 - 0:31:15) Architecture of Agentic AI Application
(0:31:15 - 0:56:49) Observability, Tracing, and Evaluation Layer
(0:56:49 - 1:11:35) Project Demo (Trailer)
(1:11:35 - 1:28:46) Break and Setup
(1:28:46 - 1:38:32) Data Ingestion Introduction (True vs Noisy Data)
(1:38:32 - 1:45:34) Chunking Strategies (Size and Overlap)
(1:45:34 - 1:49:27) Embedding Models and Vector Stores
(1:49:27 - 2:51:46) Q&A and Embedding Strategies
(2:51:46 - 3:13:40) Retrieval Pipeline and Fallback Mechanisms
(3:13:40 - 3:52:33) Deployment Discussion and Multimodal Context
(3:52:33 - 4:10:00) Ingestion Pipeline Implementation (True Data)
(4:10:00 - 4:13:23) Ingestion Pipeline Implementation (Noisy Data)
(4:13:23 - 4:21:00) Q&A and Break
(4:21:00 - 4:32:17) Retrieval Pipeline Deep Dive
(4:32:17 - 4:33:54) Reranking Models Implementation
(4:33:54 - 4:58:26) Q&A on Latency and Reranking
(4:58:26 - 5:07:00) Recap and Documentation Review
(5:07:00 - 6:11:39) Logfire Configuration and Observability Concepts
(6:11:39 - 6:25:20) Spans, Traces, and Waterfall Explained
(6:25:20 - 6:59:07) UI Integration and Prompt Security
(6:59:07 - 7:02:00) Jailbreak Protection and Guardrails
(7:02:00 - 7:13:13) Sensitive Topic Filtering and Dialogue Management
(7:13:13 - 7:21:18) Implementation of Guardrails (Topic Guard)
(7:21:18 - 7:30:12) Break and LLM Gateway Overview
(7:30:12 - 8:11:19) Trace Analysis and Error Debugging
(8:11:19 - 8:17:24) Caching, Final Q&A, and Next Steps