Next-Gen Software Architecture with Large Language Models (LLMs) like ChatGPT
summary
Demonstrates using ChatGPT (plus Perplexity, Gemini and Miro/Mirrorboard) to produce empathy maps, customer-journey diagrams, use cases, sequence diagrams and an architecture spec in minutes; generates requirements, cost estimates and nonfunctional targets (example: support 50,000 concurrent users) and compares tech choices (ElasticSearch vs OpenSearch, CloudFront vs Akamai, Kubernetes/ECS vs Lambda). Covers prompt patterns—chain-of-thought prompting, step‑by‑step, guardrails, chunking for large files, k‑NN vector retrieval and tactics to reduce hallucination (temperature=0). Outputs are draftable but need human validation.
AI-written summary · about
chapters (9)
- 0:00Introduction and talk overview
- 1:04Principles of prompting with LLMs
- 10:21Chaining prompts for better results
- 17:18Creating empathy and journey maps
- 25:56Generating requirements and use cases
- 34:27Non-functional requirements and trade-offs
- 43:59Advanced productivity techniques
- 53:20Guardrails and advanced strategies
- 59:23Summary and closing remarks
related talks
48mUsing AI in Software Design: How ChatGPT Can Help With Creating a Solution Architecture
21mChatGPT, LLMs, and LangChains - A Beginner's Guide
46mUsing AI in Software Design: How ChatGPT Can Help With Creating a Solution Arch.
56mLarge Language Models for Python Specialists
01h 00mUnleashGPT: Mastering the Implementation of Large Language Models
46mConversations with ChatGPT: illusion or reality?
33mOpening Keynote: LLMs, between fantasy and reality - the poin...
30mHow To Build Your Own Internal ChatGPT