Neelum Ghazal's Work | ContraWork by Neelum Ghazal
Neelum Ghazal

Neelum Ghazal

Full-Stack Developer | Web & Agentic AI Development

    1
    Followers
Followed by GALLERY L

Karachi, Pakistan

WorkPostsAbout

Karachi, Pakistan

Cover image for AI Customer Support Agent for
AI Customer Support Agent for Web Form, Gmail & WhatsApp An AI "digital employee" that answers customer support questions across a web form, Gmail and WhatsApp, and hands off to a human team when it should. Load-tested with 200+ messages and no lost messages. I built an AI customer support agent as a demo for a fictional SaaS company called Flowmint. Customers can write through a web form, Gmail or WhatsApp. The agent finds the answer in a knowledge base, replies on the same channel, and passes the conversation to a human team (Billing, Sales, Support) for refunds, special pricing, or when a customer asks for a person. One agent for three channels, with replies formatted for each (a clean HTML email, a short WhatsApp message, a private web tracking page) Safety rules written in code: it can only quote approved prices, and refunds always go to a human Remembers returning customers on email and WhatsApp Private ticket tracking page with a secret tracking code If the AI service is busy, it switches to backup AI models, and to a knowledge-base answer if all are limited I load-tested the agent for 24 hours with 200+ messages across all three channels, using simulated customers plus real emails and WhatsApp messages I sent myself. 212 messages in, 212 replies out, none lost Only 3.3% needed a human In a separate Kubernetes test, I stopped the worker three times while messages kept arriving, and every message was still answered exactly once Areas I'm improving: faster replies with a paid AI plan, and spam protection for the web form. Tech stack: Python, FastAPI, OpenAI Agents SDK, Kafka, PostgreSQL, Next.js, Docker, Kubernetes, Gmail API, WhatsApp Cloud API.
0
13
Cover image for AI Customer Support Agent: Web,
AI Customer Support Agent: Web, Gmail & WhatsApp A 24/7 AI customer support agent that answers customers across three channels: Gmail, WhatsApp and a web support form. It replies from the company’s knowledge base and escalates pricing, refund, legal or angry-customer cases to a human. Every customer, conversation and ticket is tracked in a PostgreSQL CRM, and on email and WhatsApp the agent remembers earlier messages from the same customer. Key features: One agent for Gmail, WhatsApp and web form Knowledge-base answers with smart human escalation Built-in CRM: customers, conversations, tickets Kafka intake, so no message is lost during restarts LLM fallback chain, plus knowledge-base replies if every model is down Duplicate-message guard and PII redaction Docker and Kubernetes deployment with auto-scaling and Prometheus monitoring Tested in a 24-hour live run: 212 messages in, 212 replies out, 0 lost, 3.3% escalation rate. Built with Python, FastAPI, OpenAI Agents SDK, Kafka, PostgreSQL, Next.js, Docker and Kubernetes.
0
23
Cover image for AI Textbook with RAG Chatbot An
AI Textbook with RAG Chatbot An AI-powered interactive textbook for humanoid robotics, with a built-in RAG chatbot that answers questions from the book’s own content. Students can highlight any passage and click “Ask about this” to get an explanation of exactly that section, or open the assistant to ask general questions. Answers stream in real time and include source citations, so readers can check where each answer came from. Key features: RAG pipeline grounded in the textbook content Selected-text mode: ask about any highlighted passage Streaming answers with source citations Light and dark mode reading experience Budget-friendly setup with local embeddings and a free-tier LLM Built with Python, React and a RAG pipeline with local embeddings.
1
0
46
Cover image for An autonomous “digital employee” that
An autonomous “digital employee” that monitors Gmail, WhatsApp and file drops, reasons about each incoming task, and writes a step-by-step action plan. Every plan goes to an Approval Center first, so a human approves or rejects it before anything happens. Once approved, the system executes the actions: creating invoices in Odoo ERP, posting to LinkedIn, Facebook, Instagram and X, sending emails, and generating CEO briefings. Key features: Gmail, WhatsApp and filesystem watchers AI reasoning loop that creates action plans Human-in-the-loop approval workflow Odoo invoicing and social media posting via browser automation Full audit logging of every action Dashboard to monitor and run everything Tech Stack: Python, MCP Servers, Playwright, Odoo ERP, and AI Agents.
0
25