Listening Post
An AI infrastructure platform that helps companies run models, tools, and agents within their own secure environment.
- Industry
- Artificial Intelligence / Enterprise SaaS
- Services
- Product Design, UX/UI, Full-stack, Branding
Context & problem
Listening Post is an AI infrastructure platform that lets companies run models, tools and agents inside their own secure environment. As Founding AI Product Engineer, I owned the product end to end: from UX/UI and the design system to frontend, backend, API integrations and production.
- AI without control — Enterprises wanted AI, but could not send sensitive data to outside services.
- Complex infrastructure — Running models, tools and agents required deep technical knowledge.
- A brand from zero — The startup needed a product, a brand and an investor story at the same time.
Process & discovery
Partnering directly with the founders, I defined product strategy and human-AI interaction, then built and shipped it myself: web, mobile, backend and deployment.
- Design — Product strategy, UX/UI and a scalable design system across product and marketing.
- Frontend — Web features in TypeScript, React and Next.js, plus mobile experiences in Flutter/Dart.
- Backend & AI — Backend on Supabase, API integrations, conversational AI and agent workflows, shipped to production.
- Infra & CI/CD — Infrastructure and GitHub Actions deployments, with AI coding agents for implementation and code review.
Results
I took the product from Figma to production and created the brand identity and investor materials for the approximately $2M seed round, backed by Desert Forge Ventures, FundNV, Awah Ventures and Kape Capital.
- ~$2M Seed Round — Investor Materials
- 0→1 Product — Design to Production
- Web & Mobile — Built End to End
- Brand Identity — Created from Zero
Tools
Design and code in one flow, from Figma to production.
- Figma — UX/UI & Design System
- TypeScript — Frontend & Backend
- React — Web Interfaces
- Next.js — Web App & APIs
- Flutter / Dart — Mobile App
- Supabase — Database
- GitHub Actions — CI/CD Deploys
- Claude Code — Implementation & Review
Learnings
- AI needs a human face — Text and voice interfaces made powerful infrastructure feel approachable.
- Own the whole path — Designing and coding the same feature removed handoffs between idea and release.
- Brand and product are one — One visual language made the startup look mature from day one.