
Zubair Khatti
AI Chatbot, Virtual Assistant Developer, Python and LangChain Expert
Compétences

Voir mes services


Portfolio
Expérience professionnelle
AI Engineer
Company âą Temps plein
Nov 2022 - May 2025 âą 2 yrs 6 mos
Xloop Digital Services - AI Engineer Built AI-powered chatbot systems, RAG pipelines, and intelligent automation tools using modern LLM frameworks and Python backend technologies. Core Expertise: âą LLM Applications & AI Chatbots âą RAG Pipelines & Vector Databases âą LangChain & LangGraph âą SQL + Document Retrieval Systems âą FastAPI Backend Development âą Multi-Agent AI Systems âą Streamlit & Gradio AI Apps Key Achievements: âą Developed multi-agent AI systems capable of SQL querying, document search, and web-based information retrieval. âą Built production-ready RAG pipelines using embeddings, ChromaDB, and LangGraph orchestration. âą Integrated Gemini and Qwen LLMs for reasoning, structured outputs, and conversational AI workflows. âą Created FastAPI backend services with JWT authentication, logging, streaming, and ML inference APIs. âą Delivered AI-powered web applications and internal tools for real-time interaction and automation. Featured Projects: DataBot -> Multi-Agent AI Assistant: âą Built an LLM-powered assistant with SQL querying, RAG document retrieval, and web search capabilities. âą Implemented LangGraph workflows, vector search, embeddings, and conversation memory. âą Added monitoring using LangSmith and deployed interactive Streamlit interfaces. SPSC / JST AI Mock Test Apps: âą Developed AI-based mock test generators using Gemini Pro and React Native. Implemented few-shot prompting using historical exam datasets. âą Optimized generation speed with parallel API orchestration and session-based exam flows. Additional Experience in Computer Vision: âą Developed AI surveillance and monitoring systems using YOLOv11, OpenCV, and InsightFace. âą Worked on face recognition, crowd monitoring, violence detection, and CCTV analytics systems. âą Managed dataset preparation, labeling, and custom object detection model training.