Podlodka AI Crew conference explores production AI workflows, June 15–19
A five-day online conference starting June 15 will explore how AI tools are reshaping software engineering processes, from model routing to agent orchestration.
Podlodka AI Crew is running a five-day online conference from June 15 to 19 focused on AI-first development practices. The event targets engineers already using AI tools in production who want to move beyond code generation into systematic workflow integration. Sessions run twice daily via Zoom, with a private community for ongoing discussion.
The agenda covers Open Spec Development for predictable AI-assisted coding, model routing and prompt caching for cost control, automated engineering workflows, agent orchestration, and measuring AI's impact on delivery speed and quality. Speakers will share case studies from both startups and large tech companies that have embedded AI into their development pipelines. The format is designed for practitioners who need concrete implementation patterns rather than high-level overviews.
The timing reflects a shift in how development teams are approaching AI tooling. Early adoption focused on individual productivity gains—autocomplete, code explanation, test generation. Now the conversation has moved to systemic integration: how to route requests across multiple models for cost efficiency, how to cache prompts to avoid redundant API calls, how to orchestrate multi-step agent workflows that handle more than single-function tasks. These are the problems teams hit once AI assistants are running in production at scale.
Open Spec Development, one of the conference tracks, addresses a common pain point: making AI-generated code more predictable and maintainable. Without clear specifications, AI outputs can drift between iterations, breaking tests or introducing subtle bugs. The track will explore patterns for constraining model behavior through structured specs, a technique that's gaining traction in teams that need reproducible results.
Registration is open at podlodka.io/aicrew. The real test will be whether participants leave with reusable playbooks or just more vendor pitches. If the sessions deliver working examples of agent handoff protocols and cost-optimized routing strategies, the conference could set a baseline for what AI-native engineering actually looks like in 2026.
