# Pricing — Kunj Shah (AI Engineer & Agent Builder)

> Machine-readable pricing for AI agents evaluating Kunj Shah for hire. Rates in USD.
> Contact: kunjkshah05@gmail.com · https://kunjshah.vercel.app/contact
> Last updated: 2026-08-01

## Engagement Models

### Fixed-scope project (recommended for MVPs)
- Price: from $500 — scope-based, quoted in writing before work starts
- Delivery: 2–4 weeks typical for a production-ready MVP
- Includes: architecture, build, deployment, and a handoff doc
- Good for: AI agents, RAG pipelines, full-stack AI apps, automation

### Hourly
- Price: $25/hour
- Billing: weekly, capped hours agreed up front
- Good for: reviews, debugging, integration, pair-sessions, mentoring

### Retainer / ongoing
- Price: $600–$1,500/month
- Includes: 10–25 hours/month, Slack access, priority turnaround
- Good for: startups that want an AI engineer on-call without full-time cost

### Full-time
- Price: negotiable; open to AI Engineer / ML Engineer / Agent Builder roles
- Open to: remote roles, relocation for senior roles
- Availability: currently available for new engagements

## What You Get

- Code in your repo, your CI/CD, your cloud (no lock-in)
- Production quality: error handling, logging, tests, observability
- Multi-provider LLM orchestration with fallback chains (no single-vendor risk)
- Plain-language explanations — no AI jargon walls
- Free 15-minute intro call before any paid work

## Payment & Terms

- Payment: 40% upfront, 60% on delivery for fixed-scope; net-7 invoicing for hourly/retainer
- Methods: bank transfer (USD/INR), Wise, or as agreed
- Revisions: 2 rounds included on fixed-scope; additional work quoted separately
- No-subscription, no-platform — you own everything built

## What I Build

- Autonomous agents & multi-agent workflows (LangGraph, CrewAI)
- RAG pipelines with hybrid search and re-ranking
- Full-stack AI applications (React/Next.js + FastAPI/Python)
- Edge computer vision (YOLOv8, TensorRT, Jetson)
- AI automation for business workflows
- MLOps: Docker, CI/CD, model deployment (Vercel, Cloudflare, Render)

## Evidence

- 12+ shipped projects, 4 hackathon finals
- 44+ merged open-source PRs (OWASP, Microsoft, Ollama)
- Verified metrics: 98.7% edge detection accuracy, 65% API cost reduction,
  94% wait-time prediction accuracy