Full-Stack AI Developer — I build MVPs, ship full-stack products, and wire the automation that runs the business behind them.

Most developers do one or the other. I do both.

📍 Rohtak, India  |  🌐 Portfolio  |  💼 LinkedIn  |  🐙 GitHub

Production-grade full-stack SaaS, AI-integrated products, and automation pipelines — tested, deployed, and ready for real users. Not prototypes. Not demos.

• Full-stack SaaS — multi-tenant architecture, OAuth, role-based access, payments (Stripe), CRM sync, Redis caching
• AI integration — Claude, Gemini, OpenAI for agentic tool-calling, RAG pipelines, AI summaries, voice agents (Vapi)
• Automation — production n8n workflows covering lead gen, CRM pipelines, missed-call recovery, AI marketing campaigns
• Testing — unit, integration, and Playwright E2E across every project. I don't deploy without coverage.

Multi-tenant SaaS unifying GA4, Google Ads, Meta, TikTok, LinkedIn, Search Console via OAuth. Claude AI summaries, Redis caching, AES-256 token encryption, role-based workspaces.
Stack: Next.js 14, PostgreSQL, Prisma, Redis, Claude API, Railway
Tests: Unit, integration, Playwright E2E (tenant isolation, token refresh, role-based UI, invite flow)
GitHub · Live

Full HR lifecycle platform — application, onboarding, performance reviews, disciplinary actions, offboarding. Multi-stage approval chains with secure single-use tokens, 3-tier RBAC, automated document generation (NDAs, contracts), CRM sync via signed webhooks, automated weekly reports.
Stack: Next.js 14, PostgreSQL, Prisma, NextAuth, Resend, GoHighLevel, Railway
Tests: 13+ Playwright E2E suites + full Vitest integration coverage across every workflow

Client-facing platform for LLC formation with Stripe checkout, document portal, compliance calendar, and stage-based ops workbench (INTAKE → DATA_QC → READY_TO_FILE → FILED → COMPLETED). Bidirectional GoHighLevel CRM sync via signed webhooks, Cloudflare R2 document storage, PDF generation.
Stack: Next.js 14, TypeScript, Prisma, PostgreSQL, Stripe, GoHighLevel, Resend, Cloudflare R2, Railway

Agentic AI platform where Claude/Gemini perform real database operations (shortlist universities, lock choices, create tasks) via validated JSON tool calling. Multi-provider AI with automatic fallback and hallucination prevention — AI cannot invent data not in the database.
Stack: Spring Boot, Next.js, PostgreSQL, JWT, Docker, Gemini API, OpenRouter, Perplexity
GitHub · Live

Unified inbox across Gmail, IMAP (Yahoo/AOL), and Outlook. Claude Sonnet powers AI summaries, 3-strategy reply drafts, and priority triage (URGENT/NORMAL/LOW/NOISE) on every sync. Full PWA with offline shell.
Stack: Next.js 14, PostgreSQL, Prisma, NextAuth, Claude API, Gmail REST API, IMAP, Railway
Tests: 80 Vitest unit tests + Playwright E2E
GitHub · Live

7-agent pipeline turning text, PDFs, images, URLs, or CSVs into a cross-linked knowledge base. Deterministic agents (intake, extraction, linking) handle zero-cost tasks; LLM agents (Claude Sonnet/Haiku) reserved for reasoning, writing, and quality validation. Semantic RAG via Google embeddings + cosine similarity.
Stack: Python, FastAPI, Claude API, Google Embeddings, Qdrant, Render
GitHub · Live

AI marketing campaign pipeline (5 connected flows), B2B lead generation (dual Apify scrapers + AI-configured targeting + email verification), missed-call recovery, GoHighLevel CRM automation, LinkedIn outreach system, RAG agent pipelines.
Stack: n8n, Claude/Gemini/OpenAI agents, GoHighLevel, Apify, Google Sheets API, Vapi, webhooks

Production-grade crawler auditing 200+ pages per run — broken links, meta tags, Core Web Vitals via PageSpeed API, JS-rendered SPA support. Background job queue, concurrent multi-site audits, PDF/CSV report export. Fully Dockerized.
Stack: Python, FastAPI, Playwright, async httpx, Docker, Google PageSpeed API

Full-stack development, AI engineering, and automation roles where I can own the build end-to-end — not be managed through it.

LinkedIn: linkedin.com/in/mohit-kdyn
Portfolio: mohit-resume20.netlify.app