🌌 Skyra: A Smart AI-Powered Investment Manager

Skyra is a next-generation multi-agent investment manager powered by GPT-4o and Retrieval-Augmented Generation (RAG). Designed to act as your autonomous portfolio team, Skyra ingests live market data, applies diverse analytical lenses, and recommends actionable steps for retail or institutional investors—Buy, Hold, or Sell—in real time.

---

🚀 Features

🔁 Real-Time Market Intelligence

• Live API integration (Twelve Data, Alpha Vantage, Financial Modeling Prep).
• Automated data transformation with Microsoft Fabric (Bronze → Silver → Gold).
• Supports real-time and batch ingestion.

🤖 Multi-Agent Reasoning (GPT-4o + Semantic Kernel)

Each agent specializes in a different investment strategy:

• 🧠 Warren Buffett Agent – Value investing logic.
• 📊 Valuation Agent – PE, DCF, Intrinsic value.
• 💬 Sentiment Agent – Reddit, Twitter, news analysis.
• 📈 Technical Agent – RSI, MACD, Bollinger Bands.
• 🧾 Fundamental Agent – Balance sheet, EPS, cash flow.
• ⚠️ Risk Manager – Risk scoring, volatility.
• 🧩 Portfolio Aggregator – Compiles and ranks decisions.

🖼️ Add image here: images/agent-flow.png
> Diagram showing all agents and how data flows between them.

🧠 RAG-Powered Insights

• Uses Azure Cognitive Search + LangChain.
• Retrieves analyst notes, filings, and documents.
• Gives context-aware decisions backed by real documents.

🖼️ Add image here: images/rag-diagram.png
> RAG architecture: embedding, search index, grounding.

🔐 Secure Credential Management

• Credentials managed by Azure Key Vault
• No hard-coded secrets – uses Managed Identity & .env encryption
• Optional support for pydantic-settings for dev profiles

🏗️ Lakehouse Architecture with Microsoft Fabric

• Full Medallion architecture:
  • Bronze: Raw API data
  • Silver: Cleaned & transformed
  • Gold: Final enriched datasets for agent use

🖼️ Add image here: images/medallion.png
> Fabric pipeline + Lakehouse ingestion pipeline.

📨 Reflex + Semantic Kernel Automation

• Autonomous Reflex trigger system powered by Semantic Kernel
• Generates AI-written summaries and email reports
• Delivery via SMTP (SendGrid or Outlook)

---

📊 Example Agent Output

Ticker: TSLA

📊 Valuation Agent: Slightly overvalued by 8%, based on DCF analysis.
💬 Sentiment Agent: Positive spike on Twitter and analyst upgrades.
📈 Technical Agent: RSI at 62, MACD bullish crossover.
⚠️ Risk Agent: Volatility increasing; watch position size.
🧩 Final Recommendation: HOLD (wait for entry point near support level).