Improving productivity with AI (tech)

Code: ENG-001 | Track: technical | Duration: 2 hr

Use sub-agents, configure shared memory in projects, create skills in repo, and understand coding agent capabilities.

Engineers adopting coding agents in a repo they can safely experiment in.

Engineering teams

Claude Code installed and authenticated; a repo participants can safely experiment in

Know your tool - silent failure modes, review habits that verify correctness, and the primitives that shape how Claude Code behaves.
Prompts that work - the essential ingredients of an effective prompt, context economy, and recovering sessions that go off track.
Sub-agents and skills - anatomy, triggers and data boundaries, with hands-on customisation of a working sub-agent.
Orchestration - running parallel agents for context compression, common topologies and their failure modes, and managing cost.
Shared memory across engineering - moving what your team teaches Claude into version control so it stops dying on individual laptops.
Wrap-up and resources - named commitments per participant and curated post-session resources.

Total time: 2 hours + optional 10-minute Q&A
Participants: Up to 15
Delivery: Remote
What you leave with: A customised sub-agent, a skill sketch and three named commitments