Peramanathan Sathyamoorthy CV & Portfolio

A modern, full-stack web application showcasing Peramanathan Sathyamoorthy's professional portfolio as a Senior Software Engineer. Built with Next.js 16 App Router, featuring profile Q&A, interactive document viewing, and automated PDF CV generation.

✨ Features

• 💬 Profile Q&A: Ask interview-style questions about experience, grounded in CV data and curated notes — see Profile Q&A architecture and src/lib/qa/README.md
• 📄 Dynamic PDF Generation: Server-side PDF creation with professional styling
• 👁️ Interactive Document Viewer: Inline PDF viewing with full browser integration
• 🔍 X search: Curated X/Twitter post search by date range (/x; /content-hub redirects here — Content Hub UI removed; see docs/content-hub-deferred.md)
• 🎨 Modern UI/UX: Beautiful shadcn/ui components with responsive design
• ⚡ Performance Optimized: Fast loading with Next.js App Router and Turbopack
• ♿ Accessibility: WCAG-oriented patterns with keyboard navigation
• 🔒 Production Ready: Cross-platform deployment support (Vercel, Netlify, AWS)
• 📱 Mobile-First: Responsive design optimized for all device sizes

🛠️ Tech Stack

Frontend

• Next.js 16 — React framework with App Router
• React 19 — UI library
• TypeScript — Type-safe JavaScript

UI & Components

• shadcn/ui — Components on Radix UI
• Radix UI — Accessible primitives
• Lucide React — Icons
• Motion — Animation (motion/react)

AI & Document Processing

• @react-pdf/renderer — PDF generation from React
• react-pdf — In-browser PDF viewing
• @huggingface/transformers — Local embeddings (Profile Q&A retrieval)

Tooling

• pnpm — Package manager (pnpm-lock.yaml, workspace supply-chain policy)
• Biome — Lint and format (replaces ESLint for this repo)
• Bun — Fast runtime for PDF generation scripts
• Playwright — E2E tests (system Brave Beta, not bundled Chromium)

📋 Prerequisites

• Node.js 24+ (LTS recommended)
• pnpm 11+ (Corepack: corepack enable pnpm)
• Brave Beta (or set BRAVE_BETA_PATH) for local E2E — see tests/e2e/README.md
• Bun (optional, for generate-pdf script)
• Git

🚀 Installation

1. Clone the repository

   git clone <repository-url>
   cd devprofile

2. Install dependencies

   pnpm install

   For stricter supply-chain installs (recommended when changing deps), use Socket Firewall or your team's wrapper, e.g. sfw pnpm install --frozen-lockfile. Workspace policy lives in pnpm-workspace.yaml.

   Optional Profile Q&A Ollama (narrative answers): copy .env.local.example → .env.local, set OLLAMA_BASE_URL, then restart pnpm dev. See tests/qa/README.md.

   Optional xAI Agentic QA Reactor: see xAI Agentic QA Reactor.

3. Sync editor profile (optional, Cursor / VS Code)

   pnpm editor:sync

   Applies .editor/profile.json → .vscode/settings.json, extension recommendations, and related editor config. Install the Biome extension, then reload the window.

4. Generate initial CV PDF (optional, first deploy)

   pnpm run generate-pdf
   # or: bun scripts/generate-pdf.tsx

🏃 Development

1. Start the dev server

   pnpm dev

   • App: http://localhost:3000
   • Hot reload via Turbopack

2. Useful routes

   • Portfolio: /
   • CV page: /cv
   • Certificates: /certificates
   • X search: /x (/content-hub redirects here; Content Hub removed — see docs/content-hub-deferred.md)
   • Profile Q&A: /qa

   Default /qa (no reactor): uses src/data/qa-index.json (git-tracked; regenerated on every pnpm build). For dev-only refresh: pnpm build-qa-index.

   • CV PDF: /cv.pdf

💬 Profile Q&A architecture

Visitors use /qa and POST /api/cv/qa. All server logic lives in src/lib/qa/ — full design, env vars, troubleshooting, and BDD tests are documented in src/lib/qa/README.md.

Big picture

One JSON contract for the UI: { answer, details[] }. Two backends, selected by ENABLE_XAI_REACTOR:

flowchart LR
  subgraph visitor [Visitor]
    QA["/qa page"]
  end

  subgraph api [API]
    Route["POST /api/cv/qa"]
    GW["handleQaRequest"]
  end

  subgraph paths [Backends]
    Local["Local-index default"]
    Agentic["Agentic optional"]
  end

  subgraph local_detail [Local-index]
    Index["qa-index.json"]
    Retrieve["BM25 + embeddings"]
    Golden["golden-routing"]
    Gen["template / Ollama"]
  end

  subgraph agentic_detail [Agentic]
    Defense["abuse-defense first"]
    Prefetch["preflight search"]
    Grok["Grok + 6 tools"]
    Coll["xAI Collections or local files"]
  end

  QA --> Route --> GW
  GW -->|reactor off or fallback| Local
  GW -->|ENABLE_XAI_REACTOR=true| Agentic
  Local --> Index --> Retrieve --> Golden --> Gen
  Agentic --> Defense --> Prefetch --> Grok
  Grok --> Coll
  Gen --> GW
  Grok --> GW

Loading

Path	When	Retrieval	Answer
Local-index	Default; also fallback if agentic fails or returns empty placeholder	Pre-built qa-index.json (hybrid RRF)	Curated golden text, CV templates, or optional Ollama
Agentic	ENABLE_XAI_REACTOR=true + XAI_API_KEY	xAI Collections API or bundled persona files	Grok (streamText + tools), with preflight search and chunk synthesis if the model is silent

Agent skills / agents: load src/lib/qa/README.md for module layout, production vs branch notes, and env tuning (XAI_MODEL, XAI_MAX_OUTPUT_TOKENS, etc.).

Optional xAI reactor setup: xAI Agentic QA Reactor below.

📜 Available Scripts

# Development
pnpm dev              # next dev --turbopack
pnpm build            # generate PDF + SW version + production build
pnpm start            # production server

# Quality
pnpm lint             # Biome lint only (errors) — best practices, correctness, unused vars/imports
pnpm lint:report      # Biome lint — full diagnostics
pnpm lint:fix         # Biome lint --write (errors only)
pnpm format           # Biome format --write (pure formatting: indent, quotes, semicolons, etc.)
pnpm imports:fix      # Organize imports only (assist action)
pnpm type-check       # tsc --noEmit

# PDF
pnpm generate-pdf     # bun scripts/generate-pdf.tsx

# Editor profile
pnpm editor:sync      # sync .editor profile → .vscode / Cursor

# E2E (Brave Beta — see tests/e2e/README.md)
pnpm test:e2e         # all projects (desktop + mobile viewport)
pnpm test:e2e:headed  # visible Brave window
pnpm test:e2e:ui      # Playwright UI (opens Brave Beta)
pnpm test:e2e:debug   # inspector + Brave

Dependency / security checks (after lockfile changes):

pnpm audit
pnpm install
pnpm type-check && pnpm lint

🧪 End-to-end tests

E2E uses Brave Beta via playwright.config.ts / playwright.brave.ts, not Playwright-downloaded Chromium.

export BRAVE_BETA_PATH=/path/to/brave-browser-beta   # optional
pnpm test:e2e

Do not run pnpm exec playwright install chromium for day-to-day work. Remove unused Playwright browsers: pnpm exec playwright uninstall.

Details: tests/e2e/README.md

💬 xAI Agentic QA Reactor (optional)

Agentic Profile Q&A on /qa using xAI Collections + Grok (flag-gated). Default /qa uses src/data/qa-index.json (git-tracked, synced on pnpm build) when the reactor is off or on fallback.

Copy .env.example → .env.local and set:

Variable	Required	Notes
XAI_API_KEY	Yes	Chat — Grok model inference for reactor answers (console.x.ai)
XAI_MANAGEMENT_API_KEY	Recommended	Collections — read-only search against XAI_PROFILE_COLLECTION (preferred over reusing the chat key)
XAI_PROFILE_COLLECTION	Yes	Collection name or ID from console.x.ai
ENABLE_XAI_REACTOR	Yes	true to enable the agentic path
XAI_MODEL	Recommended	e.g. grok-4.3 — must match a model your account can use
XAI_MAX_OUTPUT_TOKENS	Optional	Default 400 — short essence answers
XAI_REASONING_EFFORT	Optional	low (default) or high

Two keys, two roles: XAI_API_KEY is for chat (Grok). XAI_MANAGEMENT_API_KEY is for Collections (search/list). Use a read-only collections key so a leak only exposes content you already published. Upload/sync stays in console.x.ai or an external tool — not this app.

Collections setup: Create a Collection and upload profile sources (minimum: src/data/persona/ps-profile-v1.md; optional: src/data/cvdata.json, golden-qa.md, casual-qa.md, top-three-achievements.md). Set XAI_PROFILE_COLLECTION to its name or ID. Runtime reads live collection content via the Collections API on each request.

Development (with keys and collection in .env.local):

ENABLE_XAI_REACTOR=true XAI_MODEL=grok-4.3 pnpm dev --turbopack

Test production build locally (after pnpm build; same vars in .env.local):

pnpm build
ENABLE_XAI_REACTOR=true XAI_MODEL=grok-4.3 pnpm start

Deep dive: src/lib/qa/README.md (architecture diagrams, request flows, troubleshooting, tests). Also .env.example, docs/phase-1-xai-agentic-profile-qa-reactor.md.

🤖 Agent skills (AI assistants)

Portable skills for Cursor and other coding agents live under .agents/skills/. Index: AGENTS.md and .agents/README.md.

Skill	Use when
fix-dependency-security	Audit, sfw, supply-chain policy
upgrade-packages	Dependency upgrades and majors
audit-allow-builds	allowBuilds / lifecycle scripts
audit-ide-dependencies	Editor extension supply-chain
project-editor-profile	pnpm editor:sync, .editor/profile.json
react-client-expert	Client React refactors (minimal state/effects)

🔧 Building for Production

pnpm build
pnpm start

🌐 Deployment

Configured for:

• Vercel (recommended)
• Netlify
• AWS Amplify
• Any Node.js host

Vercel

1. Push to GitHub
2. Connect the repository
3. Deploy on push (use pnpm install / pnpm build in project settings)

📚 Documentation

• src/lib/qa/README.md — Profile Q&A library: dual-path architecture, env vars, production troubleshooting, BDD tests
• ARCHITECTURE.md — App structure and customization
• docs/vercel-err-require-esm-next-16.2.md — Vercel ERR_REQUIRE_ESM on Next 16.2
• CONTRIBUTING.md — Branching, Biome, React client guidelines, E2E
• CHANGELOG.md — Version history
• AGENTS.md — Agent conventions and Playwright notes
• tests/e2e/README.md — Brave Beta, headed/UI/debug modes

📄 License

This project is private and proprietary.

📞 Support

For questions or issues:

• Review the code and docs above
• See AGENTS.md and .agents/README.md for agent-oriented conventions
• Contact Peramanathan Sathyamoorthy directly

---

Built with Next.js 16 & React 19