Architecting Integrity: Responsible Software Evolution in the AI Era
summary
Claims a repeatable five‑step “AGENTS” workflow grounded in a Harvard agent framework and applies it to a CI/CD prototype with an orchestrator agent, an analyst agent, a guardian agent and runner/tasker agents; demoed behavior where known-infrastructure failures trigger an automatic rerun and high‑confidence bug candidates surface as proposed patches (humans approve). Targets: cut log‑triage time (55% report speed gains), reduce manual interventions to four or fewer per day, and balance velocity with safety via measurable metrics.
AI-written summary · about
chapters (15)
- 0:00Responsible AI introduction
- 1:25Adopting AI agents in workflows
- 3:14Speed versus trust in AI
- 5:01AI SDLC with oversight
- 6:51Agentic AI vision for business
- 8:19Responsible AI boundaries
- 10:09AGENT framework intro
- 10:53Audit phase for AI opportunities
- 13:00Gauge phase outcomes
- 13:43Engineer agentic workflows
- 16:16Navigate autonomy and risk
- 17:39Track AI system impact
- 20:12AGENT framework in practice
- 20:56Mindware updates for AI
- 21:38Q&A
related talks
35mThe Agentic AI Engineer
30mAgents Building Agents
33mThe Autonomous Enterprise
45mAI agents in practice - beyond the prototype
30mBuilding Tools with Tools: Modern AI-Driven Development
29mBuilding Trust in AI Agents
18mFrom Reactive to Resilient: Agentic AI
51mBeyond Code Generation: Governing AI Agents Across the DevSecOps Lifecycle