Agentic Development: Coding Agents, Tools & Multi-Agent Workflows
Agentic AI is shifting software development from developer-driven implementation to AI-orchestrated engineering. Coding agents now plan, modify repositories, use tools, run tests, validate outcomes, and coordinate complex workflows. After attending, engineers and architects will be able to design, supervise, and secure agentic workflows across the software development lifecycle with the right context, permissions, sandboxes, checkpoints, cost controls, and human oversight.
From AI Assistants to Secure, Cost-Aware Agentic Workflows
Learn from Industry Leaders about:
- The evolution from AI assistants to semi-autonomous and autonomous coding agents
- Agent engineering, task planning, automated decomposition, and orchestration
- Single-agent and multi-agent workflows for coding, testing, review, and documentation
- Managing context, memory, skills, prompt caching, model routing, token budgets, compression, and rate limits
- Cost-aware orchestration: selecting the right model, context window, tool path, and autonomy level for each task
- Connecting agents to tools and enterprise data through MCP, APIs, CLIs, IDEs, and internal systems
- Agent-to-agent and agent-user interaction through A2A, AG-UI, generative interfaces, and emerging interoperability protocols
- Secure agent execution through least-privilege access, non-human identities, approval checkpoints, sandboxing, network controls, and secrets isolation
- Protecting workspaces from prompt injection and unauthorized actions through tracing, kill switches, credential revocation, containment, and error recovery