A platform connecting Bangladesh's top researchers with students and professionals, offering research metrics and beginner-friendly research guidance.
Easily find any researcher by university, department, and metrics (h-index, i10-index, rank, and publication history).
Enabled advanced journal discovery, search, and compare journals worldwide by region, subject.
The "Eye of the Cosmos" project, developed by Team Artemis, is an immersive, multimedia experience designed to captivate a broad audience by harmonizing the discoveries of the James Webb Space Telescope (JWST) with compelling visuals and evocative music. In addition to providing a visually stunning journey through the cosmos, the project also includes engaging 3D educational games and interactive features to teach users of all ages about JWST and its groundbreaking mission to explore the universe.
An AI-powered, chat-driven expense tracker. Log your expenses in any language. Just type naturally and our NLP system categorizes and analyzes them instantly. Built with FastAPI, Next.js, and Hugging Face Transformers. Track your spending with zero manual hassle.
Project Link
Built using Web Speech API for native, zero-network speech recognition, WebSocket-based full-duplex communication, and an async FastAPI backend for non-blocking processing. The system streams audio directly from microphone to the Google Gemini Live API via an optimized Pipecat pipeline,eliminating transcription overhead and minimizing serialization. A Next.js + TypeScript frontend with React Hooks ensures efficient state updates and minimal re-renders. This architecture achieves sub-100ms conversational latency, supports interruption-aware, context-sensitive voice commands, real-time form manipulation, and robust recovery from network interruptions.
Project Link
A conversational AI chatbot built using LangChain, Groq, and Gemini to enable interactive engagement with documents and websites. Deployed with Streamlit for a user-friendly experience.
Project Link
CNN based image captionaning using VGG16 for feature extraction and LSTM for caption. Trained on MIMOSA dataset and deployed via streamlit.
An innovative healthcare app using a huggingface model to predict 41 diseases based on user-inputted symptoms. Integrated features for health education and proactive management, promoting preventive care
Optipulse utilizes advanced regression techniques and unsupervised learning algorithms for detecting anomalies in performance metrics, customer segmentation and satisfaction analysis. By examining key performance indicators such as latency, churn rate, packet loss, signal strength, streaming latency, and watch time, our web-based platform enables precise data analysis for making important business decisions for enterprise solutions. It also offers automated data visualizations, based on data for aiding in data-driven decision-making processes.
Project Page here!