InvoiceVision : AI-Powered Invoice Processing Platform

InvoiceVision is an enterprise-grade, AI-powered invoice processing application built on Next.js and TypeScript. It eliminates manual data entry by automatically extracting structured financial information from invoices in PDF, image, and scan formats — delivering accuracy at scale with zero manual effort.

YIME built this platform for finance teams, accounts payable departments, and enterprises that need reliable, high-throughput document intelligence with seamless integration into existing accounting workflows.

InvoiceVision : AI-Powered Invoice Processing Platform

A high-growth enterprise was processing thousands of invoices monthly — manually. Their accounts payable team was buried in data entry, prone to errors, and unable to scale. YIME built InvoiceVision, a Vision AI platform that automatically extracts, classifies, and validates invoice data from PDFs, scans, and images with near-human accuracy.

The result: 95% extraction accuracy across multi-format documents, 80% reduction in manual data entry, and end-to-end processing time cut from days to seconds — fully integrated into their existing accounting workflows.

MediCare-Bot — Your trusted health companion.

Patients and caregivers often struggle to find reliable, easy-to-understand health information without booking a doctor's appointment. YIME built MediCare-Bot — an intelligent health companion powered by LLaMA 3.3 and RAG — trained on trusted medical literature and clinical guidelines to deliver accurate, context-aware responses instantly.

The bot understands natural language queries, retrieves the most relevant content from its medical knowledge base, and responds in plain language — reducing unnecessary clinic visits and empowering users to make informed health decisions.

Affective AI — Understanding Emotions through Your Voice

Affective AI is an intelligent system designed to detect and analyze human emotions from voice inputs. By combining speech processing with advanced machine learning models, the platform identifies emotional states such as happiness, anger, sadness, and neutrality. It leverages audio feature extraction, NLP techniques, and classification models to deliver real-time emotional insights, making it useful for applications like customer sentiment analysis, mental health monitoring, and human-computer interaction.

Trend Setters - E-Commerce Conversational Chatbot

Trend Setters is an LLM-powered conversational chatbot built for e-commerce platforms to enhance user engagement and streamline customer interactions. Using Retrieval-Augmented Generation (RAG), the system provides accurate, context-aware responses by combining a large language model with a vector database of product and business data. The chatbot assists users with product discovery, personalized recommendations, order tracking, and customer support — all through natural, human-like conversations. Optimizations like model quantization ensure low latency and cost-efficient deployment.

AI-Powered Chat Assistance for Company Circulars

This system is designed to help employees quickly access, search, and understand company circulars, policies, and internal documents. By integrating OCR with large language models, the platform converts scanned documents into searchable knowledge and enables users to ask questions in natural language. Built on Microsoft Azure microservices architecture, the solution ensures scalability, secure document handling, and fast response times. The system delivers precise, context-aware answers, significantly reducing the time spent navigating lengthy internal communications.

Adult Income Prediction

This project focuses on analyzing the Adult Income dataset to understand the key demographic and socio-economic factors that influence whether an individual earns more than $50K annually. The workflow includes extensive exploratory data analysis (EDA), data preprocessing, feature engineering, and the application of multiple machine learning models for binary classification. Various algorithms such as Decision Trees and Support Vector Machines were evaluated to identify the most effective model. The project highlights how data-driven insights can be used to uncover patterns in income distribution and improve predictive performance.

TestPaper-Pro : Dynamic Test Paper Generator using Bloom's Taxonomy

TestPaperPro is an AI-powered application designed to assist educators in generating customized test papers based on Bloom’s Taxonomy. By providing subject details, syllabus topics, and sample inputs, users can automatically generate well-structured questions across different cognitive levels such as remembering, understanding, applying, analyzing, and creating. The system leverages large language models to generate multiple-choice, short-answer, and long-form questions, ensuring variety, relevance, and pedagogical alignment. Built with an intuitive Streamlit interface, the tool significantly reduces manual effort while improving the quality and diversity of assessments.