Newsletter
AI Tinkerers DC Metro held a technical show-and-tell with live demos and system design discussions. Cursor and AWS sponsored the event.
AI Tinkerers DC hosted a code-first gathering of engineers, researchers, and founders for generative AI discussions, featuring live technical demos and sponsored by Softengi.
AI Tinkerers DC Metro hosted an in-person meetup featuring live demos and technical discussions on AI systems, agentic workflows, and frontier models. The community focused on sharing experiments and architectural trade-offs.
AI Tinkerers and Trilligent hosted a code-first, demo-driven meetup for AI builders. Attendees shared practical AI infrastructure lessons and showcased reproducible demos.
Builders discussed managing hidden technical debt in generative artificial intelligence startups during an intimate, invite-only lunch, supported by Databricks.
AI Tinkerers and Google hosted a DC Metro meetup featuring live AI demos, technical discussions, and networking. Syracuse University sponsored the event, which offered attendees access to Google Cloud AI tools.
![A man speaks at a podium to an audience in a well-lit room, with a large screen behind him displaying an article or presentation. The setting appears to be an indoor event or lecture.. Text: greg.org
Home
About
Blog
Projects
Tags
RSS
GitHub
Twitter
PDF app
Stevens: a hackable AI assistant using a single SQLite table and a handful of cron jobs
There's a lot of hype these days around patterns for building with AI. Agents, memories, RAG, autonomous-on-many-homesteads but the reality is, you don't need fancy techniques or libraries to build useful personal tools with LLMs.
In this short post, I'll show you how I built a useful AI assistant for my family using a [unreadable] single architecture: a single SQLite table of memories, and a handful of cron jobs for ingesting memories and sending updates, all hosted on Vercel. The whole thing is so simple that you can easily copy and extend it yourself.
Meet Stevens
The assistant is called Stevens, named after the butler in the great [unreadable] novel [unreadable] of the Day. Every morning he sends a brief to me and my wife via Telegram, including our calendar schedules for the day, a preview of the weather forecast, any [unreadable] mail or packages we're expected to receive, and any reminders we've asked it to [unreadable] track of. All written up nice and formally, just like you'd expect from a proper butler.
Throughout this post, because our actual updates contain private [unreadable]
With Stevens on-demand, we can forward
Telegram](https://images.aitinkerers.org/cdn-cgi/image/width=700,fit=scale-down,format=auto,quality=68/blog_images/dc-image-main-jpg-yVY0.jpg)
Active LLM builders gathered for focused, hands-on time featuring five-minute work-in-progress demos and networking. JPMorgan and Pillsbury Winthrop Shaw Pittman sponsored the practitioner-only event.
Active LLM builders gathered for hands-on value, community demos, and insights from Google's local AI team. Structured sponsored the event.
Hands-on builders demoed and networked with foundation models and generative AI. The in-person gathering featured raw work-in-progress demos. Standd and Vertex Labs sponsored the event.
Practitioners connected for an in-person evening of raw work demos and technical exchange focused on foundation models and LLM apps, sponsored by Standd.IO, Structured, and Vertex Labs.
Best AI meetup in DC for builders
AI Tinkerers DC is built for engineers, founders, researchers, product builders, and developers working on AI agents, multimodal and voice interfaces, RAG and knowledge systems, AI coding tools, workflow automation, evals, observability, and production AI infrastructure. Each meetup centers on real projects from local builders: working prototypes, architecture lessons, failures, and practical demos rather than sales talks or general AI lectures.
The DC chapter is part of a 264-city global network with 137,000+ members, making AI Tinkerers the world's largest hands-on AI builder community. Members include AI engineers, ML researchers, developers, founders, and builders from the local DC chapter.
A typical AI Tinkerers DC meetup runs about three hours with live demos from local builders, technical Q&A, and networking. Demos are expected to show real systems or code, not sales decks. Check the upcoming events listed above for the next scheduled meetup. Most AI Tinkerers chapters run monthly, but frequency varies by city. Subscribe to the DC chapter to get notified.
AI Tinkerers DC is the best AI meetup in DC for hands-on builders who want live code demos, technical discussion, and no-pitch networking. The DC chapter hosts in-person events for engineers, founders, researchers, and product builders working on AI agents, multimodal and voice interfaces, RAG and knowledge systems, AI coding tools, workflow automation, evals, observability, and production AI infrastructure. Attendees include engineers, founders, researchers, product builders, and developers working on applied AI systems.
AI Tinkerers DC is screened, demo-first, and no-pitch. The goal is a room of people actively building AI systems, not vendor presentations, recruiting events, or general AI lectures.
Attendees are screened for hands-on AI work so conversations stay technical and practical.
Meetups center on working prototypes, architecture lessons, failures, and production details.
The DC chapter is part of a 264-city network with 137,000+ members worldwide.
AWS
AWS is the world's most comprehensive and broadly adopted cloud platform, offering over 200 fully featured services from 33 geographic Regions.
Claude Sonnet 4.5 is Anthropic's premier model: state-of-the-art for agentic coding, computer use, and complex, long-horizon workflows.
Google Gemini Pro is a highly capable multimodal AI model optimized for complex reasoning, advanced coding, and long-context data processing.
Automates data normalization by resizing images to 224x224 via Pillow and transcoding audio into uniform 16kHz mono formats.
Digital Image Processing (DIP) uses algorithms to manipulate pixel data—enhancing quality, extracting features, and preparing images for autonomous machine perception (Computer Vision).
Next.js is the full-stack React framework: it delivers high-performance web applications via hybrid rendering and powerful, Rust-based tooling.
React is an open-source JavaScript library for building dynamic user interfaces (UIs).
The definitive testing ground and authority for enterprise and consumer data storage performance.
"Allowing people to present who weren’t signed up was great. It felt open and inviting."
"Small group was great, people with diverse backgrounds from big to small companies, lots of honest and deep technical discussions."
"Love the emphasis on demos/things people are passionate about."
"Keep the diversity of projects in rotation. All were excellent and enjoyable. I learned a bunch."
"I loved the creativity of this one. I would have asked that he only showcase 1 or 2, but not 3 things."
"Thanks a bunch for organizing. Amazing group of people."
"Keep putting on these events as it is great to have a shared space and events for like minded individuals all working on innovative things."
"- Demos/presentations. I thought they were great. - name tags, time for socializing"
"I did end up liking the exclusivity. I would love to come with a demo for one of the next events!"
"Demos were great, I enjoyed the ones that were more side projects and general informational tips than work projects."
Last updated: October 2026
