Tech Events & Workshops

We can now generate code faster than ever, but that creates a new problem: teams can also generate confusion, duplication, and fragile architecture faster than ever. In this session, we'll explore how to design software systems that help both humans and AI move faster safely. The focus is on clear module boundaries, reusable components, stable interfaces, validation contracts, and Polylith-inspired thinking from the Clojure ecosystem. ## What We'll Cover - **Modular Architecture for AI-Assisted Development:** Structure large systems so features have clear ownership and predictable boundaries. - **Interfaces as Guardrails:** Use narrow interfaces to make code easier to replace, test, and reason about. - **Contracts & Validation:** Treat validation rules and schemas as a shared source of truth between humans, tools, and AI agents. - **Multi-Project Reuse:** Learn how shared components can support multiple products without turning into a tangled framework. - **Metaprogramming & Automation:** Use macros, code generation, and declarative patterns to reduce repetitive implementation work. ## Why Attend? If you're using AI to build software, the real question is no longer "Can it write code?" The question is: **can your architecture absorb that speed without collapsing under complexity?** This talk is for developers who want to build larger systems faster while keeping the codebase understandable, testable, and adaptable.

Join us for a practical, business-focused session on building a **voice agent** that can answer calls, collect useful information, and trigger real workflows behind the scenes. Voice AI is moving quickly from demos into real business operations. In this session, we'll look at how to connect a conversational voice layer to a backend and an event-driven workflow so your agent can do more than talk: it can capture leads, record context, publish events, and kick off follow-up actions. ## What We'll Cover - **Vapi for Voice Agents:** Build a real-time voice assistant for web or phone conversations, handle call events, and connect the agent to custom tools and backend APIs. - **Supabase as the Business Backend:** Use Postgres, Auth, Realtime, and Edge Functions to store customer context, call outcomes, transcripts, and structured lead data. - **Google Cloud Pub/Sub for Async Workflows:** Publish call events to topics so downstream services can process notifications, analytics, CRM sync, and follow-up tasks without blocking the conversation. - **End-to-End Business Flow:** Map an inbound conversation into a durable workflow: greet the caller, qualify the request, save the data, and trigger the next business action. - **Reliability & Human Handoff:** Discuss logging, retries, acknowledgements, escalation paths, and where a voice agent should stop and hand the conversation to a person. ## Session Flow - Architecture walkthrough: voice layer, backend layer, event layer - Vapi assistant setup and tool-calling strategy - Supabase schema and Edge Function integration - Publishing business events through Google Cloud Pub/Sub - Practical discussion: what makes a voice agent useful, safe, and maintainable ## Perfect For - Business owners exploring AI-powered customer workflows - Developers building voice, automation, or backend integrations - Teams interested in lead capture, appointment routing, support intake, or call automation - Anyone curious about connecting AI agents to real production systems Whether you're new to voice agents or already experimenting with AI automation, this session will give you a practical mental model for building a voice workflow that fits real business needs.