Neuropath

Neuropath is a cross-industry employee intelligence and onboarding engine. It turns a role description and an employee profile into an explainable readiness signal, prioritized skill gaps, a dependency-aware 90-day pathway, learning resources, assessments, and manager-reviewed proof.

The product is designed for more than software engineering. The seeded demo includes a technical employee, a Marketing Executive, and a manager who can review both profiles.

What the demo includes

Employee workspace

• Secure session login with three seeded profiles.
• Command center with role readiness, pathway completion, pending reviews, critical gaps, parsing confidence, and next best action.
• New analysis flow for uploading or pasting a resume/profile and job description.
• Explainable role intelligence with have, partial, and gap filters, priority levels, evidence, and a React Flow dependency map.
• 90-day pathway split into phases with effort estimates, outcomes, skill dependencies, and status controls.
• Learning library tied to detected gaps. Tavily can replace the built-in links with fresh resources.
• Evidence and feedback workspace with micro-assessments, scores, pass/fail state, and manager comments.

Manager workspace

• Team radar showing assigned employees, readiness, progress, blockers, and review status.
• Employee switcher for the technical and Marketing Executive profiles.
• Review queue for submitted evidence and assessment results.
• Approve evidence or request changes with specific feedback; the employee sees the decision in their feedback panel.

AI and resilience

• Gemini can enrich cross-industry competency extraction and role modeling.
• Tavily can curate current learning resources for high-priority gaps.
• Ollama with qwen2.5:3b can generate local assessment questions.
• Deterministic extraction, taxonomy matching, pathway generation, scoring, authentication, and review workflows remain usable when every provider is offline.

Fast local start

Prerequisites: Node.js 20+, Python 3.11+, and optionally Ollama.

1. Copy the environment template:

   Copy-Item .env.example .env

2. Start the Flask API in one terminal:

   cd backend
   python -m venv .venv
   .\.venv\Scripts\Activate.ps1
   pip install -r requirements.txt
   python -m app.main

3. Start the Next.js interface in a second terminal:

   cd frontend
   npm install
   npm run dev

4. Open http://localhost:3000.

On Windows, npm run dev clears only stale generated development files and uses Next.js Webpack mode to avoid intermittent Turbopack file-lock errors involving next-env.d.ts. Turbopack remains available with npm run dev:turbo. Activating the backend Python virtual environment does not affect the frontend.

Demo accounts

The password for all seeded accounts is Demo@123.

Profile	Email	What to show
Technical employee	employee@neuropath.ai	Engineering role fit, skill graph, 90-day pathway, assessment, proof
Marketing employee	marketing@neuropath.ai	Cross-industry competency model and non-technical resource pathway
Manager	manager@neuropath.ai	Team radar, profile switcher, review queue, approvals and change requests

The two employee profiles share the manager demo account, so one manager view demonstrates both technical and non-technical onboarding.

Optional AI configuration

Never paste API keys into source files or commit .env.

• Gemini: set GEMINI_API_KEY in the root .env. Optionally set GEMINI_MODEL to a model available to your key. Gemini is used for nuanced role modeling and competency extraction; it is not required for the core workflow.
• Tavily: set TAVILY_API_KEY in the root .env. Tavily replaces built-in learning links with current, provider-filtered resources for high-priority gaps.
• Ollama: install Ollama, run ollama pull qwen2.5:3b, keep it running, and set OLLAMA_MODEL=qwen2.5:3b if needed. Ollama is used for local quiz generation. If unavailable, deterministic assessment templates are used after a short timeout.

The app starts and remains demoable with no AI keys.

Docker

Start Docker Desktop with Linux containers, copy .env.example to .env, then run:

docker compose up --build

The compose setup builds the frontend and backend, persists SQLite in the neuropath-data volume, and connects the backend container to host Ollama through host.docker.internal.

API surface

The Flask API is session-based and the frontend sends credentials with requests.

Area	Endpoints
Auth	POST /auth/login, POST /auth/logout, GET /auth/me
Employee	GET /analyses/me, POST /analyses, PATCH /analyses/:id/progress
Assessments	POST /analyses/:id/quiz, POST /analyses/:id/quiz/submit
Manager	GET /manager/team, GET /manager/employees/:id
Reviews	POST /analyses/:id/comments

Project structure

backend/
  app/
    main.py                 Flask routes and session auth
    core/                   extraction, matching, pathway and provider services
    data/                   seeded profiles and resource fallbacks
  tests/                    backend tests
frontend/
  app/                      Next.js layout, page, styles and favicon
  components/               landing page, workspaces, React Flow and logo
  services/api.ts           typed frontend API client
old/                        local archive for PDFs and MP4s; ignored by Git
docker-compose.yml
.env.example

Verification

Backend tests:

cd backend
python -m unittest discover -s tests -v

Frontend production build:

cd frontend
npm run build

Type-check only:

npx tsc --noEmit

Presentation

Use SCRIPT.md for the two-person 2–5 minute walkthrough. It includes the order of accounts, speaker handoffs, exact screens to open, and short explanations of the AI fallback architecture.

Design and technology

• Next.js 16 + React 19 + TypeScript
• Material UI with a custom editorial dark/acid/violet visual system
• Space Grotesk display type and Manrope body type
• React Flow skill dependency visualization
• Motion-based page transitions and progressive reveals
• Flask + SQLite-backed demo data and session authentication
• Gemini, Tavily, and Ollama integrations with deterministic fallbacks