CryptoInvestYL — multi-portfolio crypto management with structured AI analysis

Context: Financial product · AI · Data
Role: Founder · Product · Architecture · Full Stack · Financial Analysis
Period: 2018 → April 2026 (main development)
Status: Still in operation — CIYL backbone
Links: Demo
White paper

In brief

Problem

Reconciling scattered data — positions, history, signals — with no room for calculation error.

Intervention

A multi-portfolio, multi-client application — consolidated data, structured AI-assisted analysis, human decisions.

Outcome

In continuous use since 2018, connected to the white-label ciyl.ch commercial offering.

Overview

CryptoInvestYL is a crypto management and trading application built like a professional product: multi-portfolio, multi-client, with a structured analysis layer — designed, architected and built end to end by someone who understands both finance and software.

The problem

Managing digital asset portfolios — your own and your clients’ — means reconciling scattered data: positions across exchanges, historical records, market signals. Existing tools handle these in silos, and the calculations leave no room for error. Every allocation, every risk threshold, every entry decision requires manual back-and-forth between sources that don’t talk to each other.

Product vision

Designed in 2018 for multi-portfolio management, the application starts from a simple principle: centralise the data, normalise it, make the analysis explainable — and keep the final decision human. The system prepares the work; the user decides.

The product then grew into an ecosystem:

CryptoInvestYL — the management and trading application, in continuous use since 2018.
CIYL — the commercial layer (ciyl.ch): the offering presented as white-label for family offices, asset managers and funds, in six languages, on sovereign Infomaniak cloud.
CIYL Engine — a content engine for SEO/AIO: articles and social media.
A public demo (demo.ciyl.ch) and a freely available white paper (github.com/ciyl/whitepaper).

Solution

One clear overview: all portfolios in the same place, with consolidated performance.
Analyses explained plainly: what weighs, what moves, why — without hunting for numbers.
Configurable alerts and dashboards drawn for quick decisions.
The decision stays human: the system prepares the work, the user decides.

Outcome

What exists today, with no invented numbers: an application in continuous use since 2018; a white-label commercial offering (ciyl.ch) in six languages on sovereign cloud; a content engine; a public white paper and a public demo. Main development wrapped up in April 2026 — the infrastructure remains in operation and serves as the technology backbone of the CIYL offering.

What this project shows

That financial logic and software can be held together — the same care runs through the portfolio model, the choice of a data source and the interface. And that in finance, reliability and explainability come before autonomy: the AI prepares, the human decides.

Under the hood

Stack, architecture decisions — the technical detail behind the outcome.

Architecture

The application is organised in clearly separated layers:

Data: collection, normalisation and caching of exchange and signal feeds.
Processing: multi-portfolio aggregation, allocation calculations, position handling.
Analysis: the structured analysis layer (see below).
Orchestration: analysis sequencing, alerts, overall consistency.
Product: multi-client management, authentication, permissions.
UI: Angular dashboards designed for fast decisions.

Data and integrations

Multi-client trading runs on the Binance API; market alerts on TradingView — two publicly documented integrations. Each one demands discipline: heterogeneous formats, rate limits, variable reliability. The answer is the same everywhere: caching strategies, retry mechanisms, systematic validation of incoming data. No incoming feed is treated as the source of truth for numbers that drive decisions.

Analysis engine

Analysis runs in three complementary layers:

Express description — Cassandra Blake, a fast read of any asset.
Fundamental analysis — Sofia Kingdom, the full layer: project, market sentiment, team and execution, market and valuation, tokenomics.
Historical timeline — Ethan Rowe, the crypto historian: the events that shaped an asset, in order.

A fourth component, Diana Ortega, holds the cache architecture — long-term, monthly, short-term — so these analyses stay fast without giving up data freshness.

AI and specialized systems — the honest classification

Cassandra Blake, Sofia Kingdom, Ethan Rowe and Diana Ortega are structured AI specialists: purpose-built prompts, AI-assisted workflows, and an infrastructure-level caching layer. They are not autonomous agents — and that is a deliberate choice.

In a financial context, autonomy is not the goal. Reliability and explainability are. The AI organises, synthesises and prepares the decision; the human makes it. Choosing structured workflows over autonomy is what makes this system defensible.

Product experience

The consolidated performance view, multi-client management and configurable alerts are the application’s daily routine. The Angular dashboards are drawn for quick decisions: check an exposure, compare a scenario, verify a signal — without hunting for numbers.

Engineering challenges

Normalisation: every exchange and signal source has its own formats, limits and incidents.
Financial consistency: multi-portfolio aggregation must be reproducible, not approximate.
Orchestration: sequencing analyses without letting the data go stale.
Security: authentication, client data confidentiality, no leaks.
Near real-time: fresh enough to decide, stable enough to trust.

My role

Founder. Product, architecture, full-stack development, financial analysis: the same thinking runs through the portfolio model, the choice of a data source and the design of a dashboard. That finance × engineering crossover is the core of this project. The product holds together software, financial logic, automation and product ownership.

Key expertise

End-to-end financial product
Financial data architecture
Binance · TradingView integrations
Structured AI-assisted analysis
Security and multi-client management
Angular front-end engineering

Today this product is commercially available as CIYL — a white-label crypto portfolio management infrastructure for asset managers, family offices and professional investment firms.

Working on a complex financial product? Let’s talk about your architecture.

Stack

Angular
Binance API
TradingView

Architecture decisions

Structured workflows over autonomy
Why : In a financial context, reliability and explainability matter more than autonomy.
Consequence : The AI specialists organise and prepare the decision; the human makes it.
No incoming feed treated as the source of truth
Why : Heterogeneous formats, rate limits, variable reliability.
Consequence : Caching strategies, retry mechanisms and systematic validation of incoming data.