At ARSSTECH18, I designed and deployed agentic workflows that process business events, use GPT-4 to reason over unstructured input, and take downstream actions across recruitment screening, chatbots, data pipelines, and fraud detection.
I integrated REST and SOAP APIs with webhook triggers and added structured output, conditional routing, retries, and fallback paths. These production workflows reduced manual operations by up to 70%.
On my Customer Support AI Agent project, I built a RAG-based system with LangGraph and FAISS to answer questions from a company knowledge base with source-grounded responses. It also routes intents, calls tools, and escalates low-confidence answers to a human.
I also built agents for call review, WhatsApp recruitment screening, and scam detection. They use tools such as Whisper, GPT-4, and APIs to classify input, update business systems, and escalate cases when needed.
