Real-Time Stock Market AI Agents

I am building a suite of AI-driven agents that track and manage Indian equities in real time. The core focus is twofold: first, deep research and data analysis that draw simultaneously on historical price archives, a live market data feed, and up-to-the-minute financial-news streams; second, an automated portfolio-management layer that can rebalance holdings, flag risk, and surface opportunities as the market moves. Here is what I need from you: • An end-to-end data pipeline that ingests those three feeds, cleans them, and serves them to downstream models with latency low enough to support continuous, real-time refreshes. • Research/analysis agents that apply modern NLP and time-series techniques (Python, Pandas, scikit-learn, PyTorch or TensorFlow are fine) to extract actionable insights from both structured price data and unstructured news text. • A portfolio-management agent that converts those insights into allocation suggestions, position sizing, and risk metrics, and that can expose its decisions via an API or dashboard. • Clear documentation and a short demonstration notebook or script so I can reproduce the workflow locally. If you can also architect the solution so predictive analytics modules can be slotted in later, that would be a plus, but the immediate milestone covers only research/analysis and portfolio-management capabilities. Please outline the approach, key libraries you propose, and any prior work with Indian market data when you reply.

Project ID: 39719449