Understand the market deeper.
What started as an idea while I was studying finance has grown into a quantitative research platform combining market data, company information, filings, financial analysis, risk models and hundreds of quantitative signals inside one evolving system.
One market can be viewed hundreds of different ways.
FinLab AI is designed around that idea. Instead of depending on one indicator, one formula or one model, the system examines market behavior, volatility, regime, fundamentals, company documents, risk and quantitative signals from many different analytical perspectives before moving toward an output.
Start with the market.
Price, momentum, volatility, regime and macro context are the surface. The system begins there — but it does not stop there.
Then go underneath the ticker.
Company filings, financial statements, company documents, earnings information and structured market data feed the research process.
0 analytical layers.
Not one signal. Not one indicator. The current prediction architecture examines the market through hundreds of analytical layers.
0 formulas, features & signals.
Standard indicators, custom variants, risk formulas, regime transforms, feature engineering and composite signals work across the system.
~0 expressions & functions.
The quantitative services underneath FinLab AI continue to expand as the platform is tested, rebuilt and refined.
A signal means little without risk.
Volatility, downside, regime, confidence, coverage and uncertainty are intended to shape how an opportunity is interpreted.
Complex underneath. Clear above.
My long-term goal is to keep the depth inside the system while making the experience above it increasingly easier to understand.
Hundreds of features can grow from a smaller set of serious financial ideas.
The 897 figure includes established market indicators as well as variants, customizations, risk calculations, regime transforms, feature engineering and composite signals. The goal is not to collect formulas for the sake of a larger number. It is to examine the same market through different horizons, transformations and conditions.
Standard indicators
RSI, MACD, ATR, moving averages, momentum measures, volatility indicators and other established market tools.
Variants & customizations
Different windows, smoothing methods, normalized forms, regime-adjusted versions and modified implementations.
Risk formulas
Volatility estimators, drawdown measures, VaR-style calculations, downside metrics and risk-adjusted return measures.
Regime & signal transforms
Trend filters, market-state detectors, scoring functions and transformations that change how signals are interpreted under different conditions.
Feature engineering
Returns across many horizons, rolling statistics, relative relationships, cross-asset ratios, fundamental transforms and news-derived features.
Composite formulas
Higher-level signals constructed by combining other formulas and features into new analytical outputs.
Building the models was only part of the problem.
Financial data is expensive, fragmented and difficult to structure well. A significant part of FinLab AI has therefore been about creating the infrastructure that allows models to work with stronger information: market data, company filings, financial statements, company documents, macroeconomic context and other structured financial sources. I wanted the system to go deeper than generic summaries and keep the underlying research connected to real financial information.
I wanted to build something of my own.
The financial markets lab became one of the places where the idea started to take shape.
I would spend time using the Bloomberg Terminal, exploring companies and markets, learning how professional financial information was organized and completing the Bloomberg training and certifications available to me. The more I learned, the more interested I became in what happened underneath the screen — how data becomes research, how research becomes a decision, and how quantitative systems interpret markets.
There was no team behind the beginning of FinLab AI.
It has been a solo project. I am a tutor at the Fowler College of Business at San Diego State University, and I have been building FinLab AI alongside my classes, tutoring responsibilities and finance studies. Some development sessions became my own version of a hackathon: define the target, set the time, and keep working until that part of the system was functioning. Many nights stretched much longer than planned. Then I would wake up and continue building.
Most of what I earned from tutoring went back into the platform.
Financial data costs money. Infrastructure costs money. Access to tools and information costs money. Building independently meant choosing where limited resources could have the greatest impact. Most of the time, I chose to put the income I earned from my part-time tutoring work back into FinLab AI.
The platform is growing alongside my own understanding of finance.
Alongside my finance degree, I am continuing my CFA journey and expanding my understanding of professional investment analysis, financial markets and quantitative finance. My goal is to continue developing experience in professional finance while continuing to build FinLab AI, so that what I learn from markets improves the platform and the process of building the platform makes me better at understanding finance.
Not as a finish line. As a measure of commitment to something I still believe can become significantly better. There are still models to validate, datasets to strengthen, interfaces to simplify and ideas waiting on the development list.
Because at some point, the platform has to meet real users.
I could continue building FinLab AI privately for another year and still find another model to add, another data source to improve or another part of the interface to rebuild. I want people to use it, question it and tell me where it can improve.
Putting it online also changes something for me: the project is no longer sitting only on my computer. People can see it. That gives me another reason to keep pushing the work forward, fix the next problem and make tomorrow's version better than today's.
See what I've been building.
FinLab AI is currently available without charge during this development and feedback stage. As the platform matures, access is expected to become paid because financial data, infrastructure and continued development have real costs. For now, explore it, test it, challenge it and tell me what should become better.
Access FinLab AI →