Data + AI Summit Relationship Agent

A small applied AI workflow app for turning conference conversations into structured follow-ups, company-fit signals, and next actions.

This project is built as a proof artifact for Forward Deployed Engineering and Applied AI roles. It shows the workflow shape: start with messy human notes, extract signal, rank the opportunity, draft a follow-up, and turn relationship context into an operating surface.

What It Does

• Captures people met at Data + AI Summit
• Detects FDE, applied AI, hiring, customer, workflow, media AI, and developer-platform signals
• Ranks contacts by priority
• Maps relationships into a follow-up pipeline
• Scores target-company fit
• Drafts concise follow-up messages
• Exports JSON, CSV, and a markdown follow-up brief
• Runs with no external dependencies

Why This Is FDE-Shaped

Forward Deployed Engineering starts in ambiguity.

The work is rarely just code. It is:

• discovery
• workflow mapping
• technical scoping
• stakeholder translation
• prototype building
• rollout
• adoption
• feedback into product

This app is a miniature version of that loop. It turns unstructured relationship notes into a workflow a person can act on.

Quick Start

npm run dev

Open:

http://localhost:4173

Run the CLI classifier:

npm run classify

View a generated sample output packet:

outputs/sample-follow-up-brief.md

Run the project check:

npm run check

Demo Flow

1. Click Load demo data.
2. Review the priority queue.
3. Open a follow-up draft.
4. Check the pipeline.
5. Copy the markdown brief.
6. Add a real contact from the Summit.

Project Shape

index.html                    Browser app
src/app.js                    Relationship classifier and UI logic
src/styles.css                Product UI
data/sample-notes.js          Browser demo data
data/sample-notes.json        CLI demo data
data/target-companies.js      Target-company fit model
outputs/sample-follow-up-brief.md  Generated sample relationship packet
scripts/classify-notes.mjs    CLI classifier
scripts/serve.mjs             Zero-dependency local server
scripts/check.mjs             Smoke check
docs/fde-product-brief.md     Product and role-positioning brief

Hiring Signal

This project is intentionally small, but it is meant to show:

• applied AI product judgment
• workflow automation thinking
• customer-facing technical language
• practical prioritization
• clean UI implementation
• ability to turn a live business moment into working software

Next Upgrades

• Add OpenAI or Anthropic API extraction for raw notes
• Add calendar/email follow-up integration
• Add LinkedIn-safe outreach templates
• Add a target-company research importer
• Add a session mode for live conference capture
• Add local file export instead of clipboard-only export

Positioning Line

I build the bridge from messy workflow to working AI/software system.