Fast and production-ready question answering in Node.js
-
Updated
Jul 18, 2023 - TypeScript
Fast and production-ready question answering in Node.js
AI-powered universal search for all your personal data, tailored just for you. Goal:The world's first product with "edge-side LLMs + consumer data localization" as its core development direction.
Using BERT for transfer learning - but just in the browser.
Use common pre-trained ML models in Deno!
AI-powered YouTube comment analysis with BERT sentiment detection, BERTopic clustering, and Ollama AI summaries. Built with Next.js 15, FastAPI, and HuggingFace transformers.
FraudSentinel is an AI-powered multi-modal fraud detection platform designed to identify modern cyber threats across emails, credentials, attachments, websites, voice, prompts, and AI agents. It combines machine learning, transformer models, audio analysis, URL and domain forensics, sret scanning, and sandboxed behavioral detection to generate.
🛡️ An AI-powered debate training and communication risk analysis platform. LogicShield strengthens your arguments, detects logical fallacies, and evaluates reputational risk before you publish, pitch, or perform.
AI-powered universal search for all your personal data, tailored just for you. Goal:The world's first product with "edge-side LLMs + consumer data localization" as its core development direction.
Price tracking, forecasting, review sentiment analysis and RAG-powered Q&A platform for Turkish e-commerce
YMM4(ゆっくりムービーメーカー4)の台詞を感情分析し、動く立ち絵の表情を自動配置するデスクトップアプリ
AI powered livestream donation filter & alert protection system. Built with a Triple Layer Shield (BERT Neural Net, Regex, and Heuristics) and real-time OBS Overlay integration to protect streamers from donation spam, gambling promotions, and malicious bots.
Turkish news classification with encoder-only BERT — PoC with full microservice pipeline, model serving, and management dashboard.
A full-stack AI-powered web app for automated news summarization and claim verification. Built as part of a final year project using React, Flask, PostgreSQL, and fine-tuned NLP models (BERT, T5).
Neutral Net is a TypeScript-React & Flask web application developed for the COS398 Ethics and Social Impacts of Computing course at the University of Southern Maine. This application utilizes SerpAPI and a fine-tuned BERT model to analyze Google News search results for bias.
ONNX-powered code embeddings for Node.js & browser — int8 quantized BERT model, zero network dependency. Semantic code search, RAG, and AI tools for TypeScript/JavaScript.
AI-powered Amazon seller intelligence platform — BERT sentiment, XGBoost return risk prediction, and RAG chatbot grounded in real customer reviews.
Development Phase! Emotract is a cutting-edge chat application that combines seamless communication with advanced features like real-time text sentiment analysis and age verification. Powered by machine learning and state-of-the-art transformer-based models.
🌼 Deliver AI-powered early ASD screening, home intervention, and pre-visit support to streamline care and empower families and clinicians.
Add a description, image, and links to the bert topic page so that developers can more easily learn about it.
To associate your repository with the bert topic, visit your repo's landing page and select "manage topics."