AI Tinkerers Portland: Building Voice Agents & Conversational AI Stacks

The bottleneck in conversational AI has shifted from simple speech recognition to robust agent orchestration. Join the Portland chapter for a deep dive into Voice AI and Voice Agents, focusing on the implementation challenges of building reliable, low-latency conversational interfaces that ship.
This is the builders’ room for engineers, researchers, and technical founders actively integrating speech-to-text, LLM inference, and text-to-speech into cohesive, production-ready systems. We are focused on the technical ‘how’—the architecture, the trade-offs, and what broke when you deployed.
Join the world’s most curated community of active AI builders on Thursday, March 5th, 2026 from 6:30 PM - 8:30 PM in Portland, OR. The specific venue address will be provided upon acceptance.

### The Signal: What We’re Building Now

We are looking for demos that tackle the hard, practical challenges surfaced by recent innovations in real-time voice interaction and agentic systems, such as:

• Low-Latency Hacks: Techniques for minimizing end-to-end latency in conversational loops, including optimized inference with models like Kimi K2.5.
• Agentic Workflows: Implementing sophisticated voice agent logic using new standards like Agent Skills Open Standard Adoption or optimizing for long-running agentic tasks with GPT-5.2 Codex API.
• Real-Time Context: Strategies for managing and retrieving conversational context efficiently for voice-driven applications.
• Implementation Deep Dives: Practical use of open-source or proprietary ASR/TTS models and the infrastructure required for scaling voice agents.

Attendance is highly curated and capped at 150 builders to ensure a high-signal environment. Space is limited, and entry is earned, not granted.

### The Stage is Reserved for Builders: Submit Your Voice AI Demo

The stage is reserved only for those actively building. We demand working code or nothing. We are looking for short, technical demos (around 5 minutes + Q&A) that expose the internals of your Voice AI stack.
Speaker Form Title: Submit Your Voice AI Demo or Talk Proposal
Description: Share your work on Voice AI or Voice Agents. We are looking for technical demos, implementation challenges, and deep dives into building conversational interfaces. Proposals should focus on ‘how’ you built it, not ‘why’ someone should use a product.
Submit Your Proposal Here

### Tentative Schedule

| Time | Focus |
| ---- | ----- |
| 6:30 PM | Doors Open & Networking (Curated Check-in) |
| 7:00 PM | Technical Demos: Voice AI & Agent Stacks |
| 8:00 PM | Q&A and Implementation Deep Dive |
| 8:30 PM | Event Concludes |

### Why Attend?

AI Tinkerers is selective by design. We screen attendees to keep conversations builder-to-builder, ensuring the room maintains a high signal-to-noise ratio. You will be trading breakthroughs with leading technical founders, researchers, and engineers from top companies (e.g., Representing major platforms like Amazon, Intel, NVIDIA, Akamai, and Okta; specialized software firms such as ServiceTitan, New Relic, and Dexcom; and cutting-edge AI startups including GPT Now, PostgresML, Digits, CaseMark AI, Transparent AI, and Innerverse AI, and more.) who are actively shipping production Voice AI systems.

> “Every meetup shaves weeks off my roadmap because someone’s already shipped the exact pattern I’m stuck on.”

If your organization ships tooling or infrastructure that accelerates builders in the voice or agent space, consider partnering with us: Sponsorship Opportunities.

***

### 📊 AI Tinkerers Portland Stats

• Attendees:
• Companies Represented: Representing major platforms like Amazon, Intel, NVIDIA, Akamai, and Okta; specialized software firms such as ServiceTitan, New Relic, and Dexcom; and cutting-edge AI startups including GPT Now, PostgresML, Digits, CaseMark AI, Transparent AI, and Innerverse AI, and more.
• Demos: We have submitted 17 demos and presented 16. The meetups have explored GPU-accelerated data pipelines, retrieval-augmented generation, multi-cloud LLM deployment, prompting and context management, AI personas and reliability, consumer AI apps, modular LLM tooling, energy-aware AI concepts, and practical lessons from open-source tools. Notable demos include Korvus, Bootstrapping products, Lac to Love, Visual prompting, and Speedrunning NanoGPT training runs.