Full-stack product work
Taking products end to end, from multi-tenant data models and admin dashboards to fast public sites that work on a phone.
- Next.js
- React
- TypeScript
- Tailwind
$ whoami
Logan Dziuk · software developer · builder
Full-stack products, from AI-powered CRMs to real estate platforms. This is where I write about what I've shipped, the tools I'm using, and the decisions I made along the way.
I build software products end to end: figuring out what people actually need, modeling the data, designing the interface, and shipping it to production.
Most of my work is in TypeScript and Next.js: recruiting CRMs, assessment platforms, and real estate sites, plus the messaging, payments, and AI integrations that make them useful. I work AI-first, both in how I build and in the features I ship, and I care about the small details that make an interface feel right.
The blog is where I write about what I've shipped, the tools I'm using, and the decisions behind them.
When I'm not working, I'm skiing in the winter and traveling in the summer.
export const logan = {
role: "software developer & builder",
focus: ["product", "AI features", "integrations"],
stack: ["TypeScript", "Next.js", "Postgres"],
offline: ["skiing", "traveling"],
motto: "obsessively crafted",
} as const;
// TODO: write more of this downTaking products end to end, from multi-tenant data models and admin dashboards to fast public sites that work on a phone.
Agents that call tools with guardrails, AI-written summaries and next steps, and voice, all routed through the AI SDK and AI Gateway.
Texts, calls, and email, from one-off messages to scheduled digests and campaigns, plus public REST APIs with signed webhooks.
Postgres and Firestore data models, sign-in, and scheduled jobs, deployed to Vercel with automated checks on every merge.
The tools I use day to day, plus a few I pull in when a project needs them.
A recruiting CRM for real estate brokerages. Texts, calls, email, and video share one inbox, an AI agent drafts follow-ups and summarizes prospects, and automations run the nurturing. Teams plug in through a REST API and webhooks.
Shows brokerages how agents like to work before sponsoring them. Agents take a branded assessment, get scored against a role profile, and receive an AI-written coaching report. Hiring teams connect through a public API and signed webhooks.
Our company site, rebuilt from WordPress in Next.js. Every legacy URL 308s to its new home so search rankings carried over, the blog archive imports in one click, and every form lands as a lead in a built-in CRM.
A site for a Twin Falls REALTOR®: live MLS search on a map, saved searches that email new listings and price drops, and an admin with a built-in CRM and campaign editor.
A site for a Twin Falls home restoration and investment business: a portfolio of restored homes, lead forms for sellers and investors, and an admin for projects and photos.
A Jackbox-style Family Feud: the board runs on a TV and everyone buzzes in from their phone, synced in real time. An optional AI host writes surveys on any topic, judges answers, and talks back in its own voice.
A light, fast CRM for small B2B teams: pipeline boards, one timeline per record for calls, emails, texts, and tasks, and branded email from shared inboxes to campaigns. MIT licensed, and runs locally with zero config.
My personal AI agent as a deployable workspace: it works in a sandboxed shell over its own private files, remembers through notes and embeddings, runs scheduled jobs, and triages its own email inbox.