What you’ll learn
Most AI-coding content is “I built a todo app with Claude”. This course is the opposite end — serious engineering on real artifacts, applied over months, with 90+ executed plans, 1000+ tests, and a working compiler behind it.
- Plan-driven autonomous execution — how to write a plan an agent can execute end-to-end without your daily babysitting.
- The audit cycle — a second agent with a different framing that catches assumption holes your tests miss.
- Worktree discipline — isolating agent work so parallel tasks don’t collide.
- Reviewing AI-generated code — patterns for fast, reliable review at scale.
- Failure modes at scale — hallucinated facts, defending past positions, off-distribution, echo chamber. How to recognise and mitigate each.
- Threshold-based escalation — what to delegate, what to keep, and where the line moves over time.
Who it’s for
Middle and senior engineers, tech leads, and engineering managers who already use AI assistants and want to move from “helpful autocomplete” to “delegated execution”. Comfort with Git, code review, and architectural decisions is assumed.
Format & logistics
- 4–6 weeks, cohort-based.
- Live sessions + recorded materials.
- Real homework: each module ships with a concrete task on a real codebase.
- Office hours: weekly Q&A with the author.
- Pricing: early-bird from the waitlist; full price after launch.
- Language: bilingual content (RU + EN); live sessions in the participant’s language.
Join the waitlist
No spam. You’ll get a notification when the first cohort opens. Early-bird pricing is reserved for waitlist members.
Questions? Email hello@nv-lang.org.