Mark Stolan

About

I'm a product leader and hands-on builder working at the intersection of AI and the problems practitioners actually face. Over the past year I've shipped three production AI applications as the sole product owner and builder, working primarily through Lovable, Claude Code, Supabase, and Railway. My background is a decade in K-12 education as a math teacher, assistant principal, and instructional coach, which means I bring both the technical fluency to build and the practitioner context to know whether what gets built will actually be used. I'm based in Roseville, CA and currently open to senior AI product roles.

Projects

NTN AI Instructional Navigator

Teachers were feeding NTN's instructional resources into general-purpose AI tools and getting inconsistent, unreliable results because the tools had no domain context and no way to evaluate whether the output was actually correct.

A live multi-agent AI application that helps teachers design structured project-based lessons using NTN's instructional library. The tool moves teachers through a structured workflow rather than a blank prompt box, with AI generating content that has been tested against expert standards before it reaches a user. Includes a RAG pipeline with 122,785 vectorized academic standards across all 50 states, Google Drive export, and a dedicated admin dashboard for monitoring usage and iterating on output quality.

Stack Lovable Railway Supabase Google Gemini Pro/Flash Vercel GitHub

Built from concept to production in under four months. Currently in Phase 2 beta with active teacher users.

See a sample lesson

Sample lesson output will appear here. Paste a complete lesson from your own Navigator testing, with any identifying details removed. Plain text or lightly formatted text works well here.

Echo Export Tool

NTN staff needed to export hundreds of LMS courses before a platform sunset deadline, and no existing tool could handle the volume at scale.

A full internal SaaS application with team authentication, CSV upload, async batch job processing, live per-course progress tracking, ZIP download, and auto-cleanup. Team members log in and monitor jobs in real time without ever handling underlying system credentials directly.

Stack Express EJS Railway GitHub

Deployed and in active use by NTN staff. Reduced a months-long manual process to a managed batch operation.

See a sample export summary

Sample export summary will appear here. Paste a job completion summary or export report, with any course names or staff details removed.

NTN Facilitation Thought Partner

Instructional coaches spend significant time planning professional development sessions from scratch, despite drawing on a consistent set of facilitation frameworks and organizational knowledge each time.

A multi-agent AI workflow that functions as a genuine thought partner for coaches planning PD sessions. The user brings a facilitation problem and works through a structured process to leave with a plan they helped shape, not a block of AI-generated content to edit. Built following structured SME review and comparative analysis of a competing tool to validate the approach before committing to the architecture.

Stack Lovable Railway Supabase n8n GitHub

Currently in alpha testing with NTN coaches. Feedback from the initial session is informing the next build phase.

See a sample PD session plan

Sample PD session plan will appear here. Paste a complete or near-complete output from your alpha testing session, with any identifying details removed.