Megha User

I’m a Data Science professional with around 1 year of experience at Zigment, specializing in building and deploying Agentic AI solutions for customer-facing applications. I focus on delivering accurate, context-aware responses using strong machine learning and NLP foundations.

In my current role as a Forward Deploy Engineer, I built and deployed 2 agentic AI solutions for a premium jewellery brand, handling around 500 customer queries per day. I designed 5 end-to-end conversation flows, executed around 100 QA test cases, and iteratively optimized prompts—reducing response errors below 5% and achieving an 85% user satisfaction score.

I’m proficient in Python, machine learning, and natural language processing (NLP), with hands-on experience in conversation flow design, prompt engineering, QA testing, and performance optimization. I use NumPy, Pandas, Matplotlib, and Seaborn for data work, and I’m comfortable working across common cloud and data tools like AWS (including EC2 and S3).

I also back my work with research and projects, including a publication on crime prediction and forecasting using machine learning and deep learning techniques, where XGBoost achieved an accuracy score of 82%. My final year project focused on crime prediction using ML/DL (XGBoost accuracy: 82%), and I’ve presented an AI chatbot project using Seq2Seq with Turing Test in addition to earning best presentation at UDYUKTA 2K19.