Vanshdeep S.

Most calls, leads and follow-ups still depend on someone remembering to do them. I build AI agents and voice agents that handle that work. They pick up the phone, book the appointment, find answers in your company data, and update your CRM when the call is over.

$100K+ earned on Upwork | 4,300+ hours | 10+ years in software

I'm a full-stack developer who focuses on AI automation. I build the agent and also the app, backend and integrations around it, so you deal with one person instead of three. Most of my contracts run for months.

WHAT I BUILD

Voice agents
For support and sales calls. Most of the effort goes into what makes a call feel natural: quick replies, people interrupting mid-sentence, and a clean handoff to a human when the agent should stop. They connect to GoHighLevel, HubSpot or Salesforce, so the result of every call lands in your CRM.

Agent workflows
Several AI agents sharing one job, for example one researching, one writing and one carrying out the steps, built with LangGraph or CrewAI. Anything important waits for a person to approve it.

RAG assistants
Chatbots and assistants that answer from your own documents and database, so replies come from your data and not from the model's guess.

Automation
n8n and Zapier workflows that act on the result: book the slot, send the follow-up, move the lead to the next stage.

Full-stack development
The product around the AI: web apps, dashboards, APIs and databases. React and Next.js on the frontend, Python (FastAPI) and Node.js on the backend, PostgreSQL for data, Stripe for payments. I deploy to AWS, Vercel or Railway.

PRODUCTS I'VE WORKED ON

Voice and agents

• Hume.ai and Tabbly: AI voice agents
• Jivi: medical AI
• Impress.ai: AI agent for HR management
• Piktochart AI Instagram post generator: AI agent for social media content
• Copy.ai: AI for advertising and marketing

RAG and chat

• TaskingAI, ChatbotKit's art chatbot, and ToolsADay's FAQ generator: RAG-based chatbots and tools
• Raccoon AI: AI chat

Full-stack AI products