Mohsin Ahmed - AI Agent Engineer | ContraWork by Mohsin Ahmed
Mohsin Ahmed

Mohsin Ahmed

n8n + AI Automation | Save hours with smart workflows

New to Contra

Mohsin is building their profile!

Followed by Johnson A
Cover image for I've been working on an
I've been working on an AI Customer Support Chatbot for a fictional fashion store, FashionHub, designed to automate customer support while keeping human assistance available when needed. πŸ› οΈ Tech stack: n8n β€” Workflow automation WhatsApp Cloud API β€” Customer messaging Groq LLM β€” AI-powered responses RAG + Google Docs β€” Knowledge-based answers Gemini Embeddings + Vector Store β€” Knowledge retrieval Airtable β€” Conversation and order management Telegram β€” Human handoff and error alerts ✨ Key features: βœ… AI-powered customer support βœ… Retrieval-Augmented Generation (RAG) βœ… Conversation memory βœ… Order processing with customer confirmation βœ… Human handoff for complex issues βœ… Telegram notifications for support and workflow errors βœ… Webhook-based message processing πŸ’‘ My biggest takeaway: Building an AI agent is not just about connecting an LLM. It’s also about designing the workflow around it β€” handling data, retrieving reliable information, managing conversations, and planning for errors and human intervention. This is a portfolio project built for hands-on learning, and I’m continuing to improve its reliability and production readiness. I’m open to opportunities involving n8n automation, AI agents, API integrations, and workflow implementation. If you’re building AI-powered business workflows, I’d love to connect and learn from your experience!
2
0
26
Cover image for Built an AI-powered Lead Qualification
Built an AI-powered Lead Qualification & Sales Automation System using n8n that automatically scores, routes, and follows up on incoming leads β€” eliminating manual lead triage. Key features: Automated data collection with validation and duplicate removal (email + record-level checks) AI-based lead qualification using a dual-model setup (Groq + Google Gemini as fallback) for reliable classification even if one AI provider fails Smart routing based on lead temperature β€” Hot, Warm, and Cold β€” each with a different follow-up path Hot leads: instantly assigned to a sales person, internal manager alert sent, and CRM record auto-created Warm leads: logged and saved for the team to follow up, with a Slack notification Cold leads: archived to a sheet for future nurturing campaigns Automated personalized auto-reply email sent to the lead within seconds Built-in error handling β€” if AI qualification fails, the system alerts the team via Telegram instead of silently dropping the lead The result: no lead falls through the cracks, sales reps only spend time on qualified opportunities, and response time drops from hours to seconds. Tech stack: n8n, Groq, Google Gemini, Gmail, Slack, Telegram, Google Sheets, CRM integration.
0
29
Cover image for Built an enterprise-grade AI customer
Built an enterprise-grade AI customer support ticket automation system using n8n and Google Gemini, designed for zero ticket loss even during API outages. Key features: Real-time ticket ingestion via Webhooks with automated data cleaning Idempotency check against Airtable CRM to prevent duplicate tickets and redundant costs AI-powered routing β€” Gemini classifies department, sentiment, and priority Built-in fallback logic: if the AI API goes down or hits rate limits, the system automatically defaults to safe routing (General/Medium priority) instead of failing β€” the workflow never stops Dual-data sync: tickets saved to Airtable, backed up in Google Sheets, and pushed to Slack for instant team alerts Automated customer response loop via Gmail The result: a fully autonomous support system that handles volume spikes, filters duplicates, self-heals during AI outages, and keeps operations running with zero manual intervention. Tech stack: n8n, Google Gemini, Airtable, Google Sheets, Slack, Gmail, Webhooks, JavaScript.
0
37
Cover image for If your team is manually
If your team is manually copy-pasting data between tools (or dealing with sync headaches), this kind of setup usually pays for itself within weeks. Happy to share how it could work for your specific tools β€” just DM me.
0
40