About Feronix
AI-driven research. Independently tested.
Feronix is a financial research platform at the intersection of AI, machine learning and statistics. We investigate what company announcements and market data reveal—and test whether that information has predictive value.
Founder
Floris Kieviet
Founder & Research Lead
Floris Kieviet is the founder of Feronix. Driven by a deep curiosity about financial markets, AI and decision-making, he is building a research platform that makes complex financial information understandable, testable and publicly accessible.
As a young, driven founder, Floris combines an entrepreneurial outlook with a critical research mindset. He believes innovative models become valuable only when assumptions are visible, results are independently tested and disappointing outcomes are reported honestly too.
Through Feronix, he aims to bridge the gap between advanced data analysis and clear, responsibly supported insights—without exaggerated promises and with careful attention to methodology, uncertainty and context.
Built with AI
From idea to platform through vibe coding.
AI-assisted development: build faster, keep testing critically.
Feronix’s website and research software are developed with AI assistants. This approach is also known as vibe coding: ideas are expressed in everyday language and translated into prototypes, code and interactive applications with AI assistance.
AI accelerates development. It does not replace source verification, software testing or independent evaluation of research results. Floris Kieviet remains responsible for the research decisions and everything Feronix publishes.
There is an important distinction: AI assistants help build the platform, while Qwen3-4B LLM analyses financial announcements. Predictive value is tested separately.
Explore AI’s role in the architectureOur position
Research before conviction.
Financial markets contain many patterns, but few survive beyond the period in which they were discovered. Feronix was founded on one simple principle: a model earns trust only after surviving a fair attempt to disprove it.
We work with public market data, official company documents and transparent evaluation methods. Our publications provide not just a headline figure, but also the study design, comparison baseline, example predictions and limitations.
Feronix does not sell trading advice or present historical outcomes as certainties. Our aim is to distinguish more clearly between an interesting indication and a reproducible result.
What we work on
Four research areas.
Market structure
Short- and long-term patterns in prices, returns, trading times and relationships between markets.
Corporate communication
Information in earnings releases, guidance, risk disclosures and changes in management language.
Decision data
Public transactions and events that may reveal new information about expectations.
Model validation
Understanding when a predictive result generalises—and when complexity mainly adds a false sense of certainty.
Our commitment
We publish what the data shows, not what sounds most appealing.
That means absent effects, unsuccessful replications and practical limitations belong in the research archive too.
Transparency
What visitors can expect from us.
-
No hidden final test
The separation between development and evaluation is explicitly stated.
-
No claims of certainty
Findings are presented with uncertainty intervals and the scope of the data.
-
Not investment advice
This website is a research archive, not a recommendation to buy or sell securities.