KM Kanav Mehta Quantitative Researcher

QuantFlow

An AI workbench for equity research. Research a stock in chat, run the analysis in notebooks, or open Muse CLI sessions in the same workspace. Agents call quant tools over MCP: alpha_engine for prices, factors, backtests, attribution and portfolio optimization; KizoNLP for financial NLP.

Built on ByteDance's DeerFlow 2.0. Eleven specialist agents gather evidence and argue the bull and bear case; a risk manager can veto the call before the report ships. Research runs on point-in-time S&P 500 membership.

Back to Projects View the equity engine on GitHub

How the research works

  • The data engine uses S&P 500 membership as of each research date, back to 1996. That keeps stocks that later delisted in the historical universe.
  • Specialists gather evidence in parallel. The bull and bear cases are evaluated independently, with downgrades documented and a risk manager able to veto the recommendation.
  • Automated checks verify as-of dates before delivery.

Amazon chart examples

Amazon adjusted closing price, 50-day and 200-day moving averages, and trading volume
Price history with moving averages, trading volume, and 52-week highs and lows.
Amazon price history with support and resistance levels marked by horizontal lines
Reference levels plotted against the historical price series.