.png&w=256&q=75)
Qazi Tabish Firoz Ahmed@tabishqazi20
1
Events attended
1
Submissions made
India
4+ years of experience
About me
Leading AI and software development at AI Dynamics Automations, where I design and deploy LLM-powered applications, agentic AI systems, and automation workflows using OpenAI, LangChain, and CrewAI. My daily work involves architecting scalable backends in FastAPI and Node.js, integrating vector databases like FAISS and Pinecone, implementing CI/CD pipelines, and collaborating across teams to deliver production-grade SaaS and AI solutions.
🤓 Latest Submissions
.png&w=828&q=75)
D2C Ops Autopilot
D2C brands lose money in the gaps between departments. Finance does not know what quality just found. The team that handles reordering does not know about a defect until it is too late. D2C Ops Autopilot closes that gap. Six specialist AI agents work together inside a shared command room, the same kind of chat space your team already uses, to run real day to day operations: customer returns, inventory reordering, quality investigations, delivery problems, and daily financial reporting. A watcher agent sits in that room as the supervisor. People type plain requests, no special syntax, no new tool to learn. The watcher decides whether to answer right away or to open a dedicated workspace and bring in the right specialists, support, operations, finance, and escalation, who then work the problem together and hand the result back. The centerpiece is cross-department awareness. The system keeps track of every piece of work in progress, by product, order, customer, or location. When a new request touches something already under review elsewhere, the system catches it. In our flagship example, a quality investigation is open on a product with a known defect. When a request comes in to reorder that exact product, the system recognizes the conflict and holds the purchase order before any money moves, preventing a real mistake that no single department would have caught alone. The system also answers situational questions directly, "what needs my attention right now," and gives a short, ranked answer pulled from live information, without starting anything new unless asked to. Everything runs on realistic test data today. Every connection to a real business system, orders, accounting, customer messaging, is built so that switching to live data is a simple settings change, not a rewrite. This is built for D2C brands running a few hundred orders a month and up, where manual coordination between teams costs real hours every week, and occasionally costs a lot more.
19 Jun 2026
