.png&w=256&q=75)
2
1
India
1 year of experience
AI and Machine Learning practitioner with a foundation in Computer Science and Engineering, focused on data-driven modeling, intelligent systems, and applied machine learning. I enjoy working on end-to-end ML pipelines, experimentation, and practical AI solutions, and actively participate in hackathons and collaborative projects to strengthen my skills.
.png&w=828&q=75)
Qubic Guardian is an AI-driven real-time on-chain monitoring system built for the Qubic ecosystem. It continuously: - Listens to live QX asset trades - Enriches trades with real-time price data - Performs anomaly detection - Assigns risk scores + tags - Stores clean data in Google Sheets - Sends Telegram alerts for high-risk events - Updates a real-time Streamlit dashboard This system emulates exchange-grade surveillance engines, but using decentralized Qubic data + AI. ## 🔍 What Qubic Guardian Detects - **Suspicious Transactions** - **Whale Movements** - **Token-Specific Trading Behavior** - **Category-Level Risks** - **Liquidity Anomalies** - **Historic & Real-Time Trends** --- ## 🧠 Key Features ### 1️⃣ Real-Time Transaction Ingestion - QX trades are streamed through **Qubic EasyConnect Webhooks**. - Every trade instantly flows into **n8n** for processing. --- ### 2️⃣ Live Price Enrichment - Uses **MEXC QUBIC/USDT price API** to calculate real USD impact. - Enables accurate **risk scoring** based on trade size and value. --- ### 3️⃣ AI-Powered Summaries LLM (Gemini 2 Flash / DeepSeek / Groq) generates: - Human-readable explanation - Risk interpretation - Recommendation: **Monitor / Caution / Urgent** --- ### 4️⃣ Multi-Layer Risk Engine #### **A. Base Risk (Features & Risk Node)** Risk based on `trade_value_usdt`: - **< $50 → LOW** - **$50–$500 → MEDIUM** - **> $500 → HIGH** #### **B. Anomaly Detection Engine** Analyzes multiple signals: - Trade size - Shares - Whale thresholds - Illiquid assets - New / experimental tokens - Selling pressure - Price/volume irregularities - Other rule-based anomaly flags and other features.
7 Dec 2025