Our Approach

Our stock ratings methodology and portfolio construction approach are built on a four-step, end-to-end investment process designed to operate at scale.

wave

1. Data Ingestion

Data is the foundation of our platform. We ingest massive volumes of raw market and regulatory data directly from primary sources such as the SEC, rather than relying on pre-cleaned third-party datasets. Regulatory filings are inherently noisy, inconsistent, and error-prone, requiring extensive in-house cleaning, validation, and normalization before they can be used reliably. Market price data is sourced via subscription-based APIs and processed into standardized, high-quality time-series datasets.

feature
feature

2. Research and Analysis

Our research and analysis layer operates on more than 15 years of historical market and financial data at scale. We compute and analyse hundreds of financial ratios, multiples, and derived features per stock across our entire coverage universe, resulting in over a billion individual data points. This process requires production-grade computing infrastructure, including large-memory systems and hardware-accelerated workloads, to operate reliably and reproducibly.

3. AI Ratings

We apply modern machine learning techniques to transform this high-dimensional data into actionable stock ratings. Each stock is evaluated across multiple factors—including total, return, growth, risk, valuation, and technical models trained on large historical datasets. Ratings are recalculated daily to reflect new data and evolving market conditions.

feature
feature

4. Portfolio Construction

In the portfolio construction step, we translate stock ratings into diversified, rules-based portfolios. We offer predefined portfolios built around specific rating factors, as well as advanced screening tools that allow account holders to construct and manage their own portfolios using our data, ratings, and constraints.