The Homebrew Computer Club for the AI Age

On Thursday, April 16th, AI Tinkerers Austin returns to the Omni Hotel for an evening dedicated to the technical implementation of generative AI. This is a practitioner-only gathering focused on the “how” behind the build. We prioritize raw demos, messy experiments, and architectural deep-dives over polished pitches.

As the AI stack shifts toward agentic reliability and high-efficiency local execution—driven by recent breakthroughs like GPT-5.4’s mid-thought steering and 1-bit quantization models—the gap between a prototype and a production-ready system is widening. This event is where we bridge that gap by sharing what actually works in the IDE.

Event Details

• Date: Thursday, April 16, 2026
• Time: 6:00 PM – 9:00 PM
• Location: Station Austin - 701 Brazos Street Austin, Texas 78701
  Room: Captain America (8th Floor)
• Access: Strictly limited to active builders; registration requires technical screening.

📢 Call for Demos: Show Your Stack

We are looking for 5-minute technical demos that pop the hood on your current projects. We want to see your code, your agentic workflows, and your context-engineering setups. Whether it’s a novel implementation of OpenClaw for orchestration or a custom quantization pipeline for local LLMs, we value technical discovery over product marketing.

Submit Your Demo Proposal

Note: Demo presenters receive priority admission.

Schedule

• 6:00 PM: Doors Open & Technical Networking
• 6:45 PM: Kickoff & Community Update
• 7:00 PM: Technical Demos (Live code, no slides, deep implementation notes)
• 8:15 PM: Peer-to-Peer Breakouts & Networking
• 9:00 PM: Event Close

Venue Sponsor

Station Austin: “Thank you to Station Austin for sponsoring AI Tinkerers. Station Austin is the center of gravity for entrepreneurs in Texas. They bring together the best entrepreneurs in the state and connect them with their first investors, employees, mentors, and customers. To sign up for a Station Austin membership, click here.”

Curation & Community Standards

AI Tinkerers is a community for those actively shipping code. Attendance is restricted to engineers, researchers, and technical founders. We maintain a high-signal environment by screening all applicants for demonstrable technical work. This is not a venue for recruiters, marketers, or consultants.

Interested in supporting the builder ecosystem? View our sponsorship opportunities.

AI Tinkerers Austin Stats

• Community Reach: This community of 891 technical professionals features a high concentration of AI and machine learning expertise, with 82% specializing in LLMs and RAG. The membership includes senior leaders from Google DeepMind, Tesla, and Amazon. Notable for its focus on agentic workflows and autonomous systems, the group bridges the gap between theoretical research and production-grade AI deployment.
• Organizations: AI leaders and engineers from tech giants like Google, Amazon, Microsoft, and Nvidia, alongside teams from xAI, Adobe, Databricks, and Tesla, and emerging startups including Torc Robotics, Comet, ClosedLoop.ai, Loman AI, and more.
• Technical Depth: 43 demos have been submitted and 40 have been presented. The most exciting themes have focused on agentic AI for real workflows, RAG/hybrid search for production-grade retrieval, and LLM engineering practices like eval/optimization, orchestration cost controls, and scalable LMOps. Highlights include Russell Sadrieff’s AI PM workflow, Julian Ghadially’s DSPy optimization experiments, J. Michael Rozmus’s Bedrock RAG chatbot, and Michael Samon’s legislative monitoring system.
• Builder Feedback:

  “...this was an incredible resource and finding the event within ~36 hours of moving is great! I wanted to offer feedback that the marketing / ease of sign up / process was incredible and just give that positive feedback.”

  — Startup cofounder

  “Well organized event, Thank you Achyut and AI Tinkerers, Sapphire ventures for hosting the event... Venue was great and welcoming, ample space, well hosted and food and drinks”

  — Founder, Health AI

  “Great technical talks. Continue the focus on technical implementation!”

  — VP of Data, fintech

  “...i enjoyed project building time and collaboration with my team members... we were able to build a working project and show the demo, and won the first prize”

  — Founder, Health AI

  “Great work, and I'm glad to learn! Your research summary and examples were excellent!”

  — LPC & AI engineer

  “...I loved the conversation and the mix of talent that attended the event.”

  — VP of Data, fintech

  “Loved the event. I think we should have a component of the conversation that is more structured.”

  — VP of Data, fintech

A great demo (per the strongest-rated examples and the audience’s requested improvements) feels like an interactive system you can follow end-to-end: show the agent’s workflow in action, not just the concept. Aim to turn output into visible artifacts (documents, reports, structured graphs, action execution traces) and make the architecture legible through concrete mechanics (protocols like MCP, measurable optimization results, or specific implementation components). Provide at least one realistic use case that exercises the “frontier” behavior (agent/tool loop, dynamic interface generation, or actionable integrations) and ensure the audience can learn by watching what the system does—this aligns with the form’s emphasis on functional prototypes that move beyond text-based chat into agentic, interactive interfaces. Avoid demos that are primarily high-level descriptions without enough implementation detail or proof-of-function, and avoid relying on the audience to infer the experience—explicitly include a live demonstration or example product/launch scenario and a short example use-case summary to reduce the “so what does it do?” gap noted in feedback.

In Austin, DSPy and Prompt Optimization Experiments by Julian Ghadially was highlighted as “excellent” and “great work” because it shares research-backed experimentation on DSPy prompt optimization applied to a fact-checker, including strong evidence-based framing (measurable performance improvement) and clear modular system components. In Austin, GitKB – A distributed knowledge base protocol for agentic engineering at global scale by Matt Walters stands out because the audience strongly endorses the concept’s practical architecture: it’s a local-first, git-like knowledge graph protocol with sparse sync semantics, and the talk justifies why that matters for agentic coding velocity. In Austin, ClawForce: Let Claude Spawn People to Do Things at Scale by Mike Angstadt is liked for enabling physical-world workflows from an AI agent via MCP-based orchestration and gig-economy “human provisioning,” with the demo centered on simple one-line commands that “just show up” as real-world actions and measurable proof-of-work examples. Finally, in Austin, Building AI Product Manager. Using AI to help PMs move faster and enable engineers/founders to do product work like a pro. by Russell Sadrieff received a perfect rating, and while the only explicit feedback called for more live demonstration/use-case depth, the praise itself implies attendees valued the operational “quality standards” framing and the fact that the speaker built the app from scratch and can talk through the building process and the intended product workflow.