Tareluntik - abstract visualization of data flows and analytical models
Predictive analytics for remote investors

Intelligence that travels with you

Tareluntik combines algorithmic decision making with historically proven strategies, so you can manage investments from any time zone, without constantly monitoring the screen.

Get started now

Signals derived from massive datasets, not gut feeling

The platform continuously processes market, macroeconomic and business data and converts them into concrete recommendations for action. Each signal is backed by statistical analysis, not prediction based on current market sentiment.

Instead of a single general model, Tareluntik uses a set of specialized forecasting models that are benchmarked against each other and calibrated against market conditions. The result is a recommendation that considers several scenarios at once, not just one.

  • Real-time processing Data streams are analyzed on the fly, without waiting for a daily or weekly update.
  • Multi-layered signal verification Each recommendation is compared to several independent models before it reaches your supervisory review.
  • Traceability of decisions Each proposal includes an explanation of the data base on which it is based.

Three fundamental stages on which every recommendation is based

The system does not operate on assumptions. Each step is verifiable and documented so you can understand why the recommendation was designed the way it is.

01

Data collection

The system combines market, macroeconomic and sector data from multiple sources into a single, structured set suitable for further analysis.

02

Predictive modeling

A set of statistical and machine-learned models estimate the likelihood of different outcomes, weighting the results against each other based on reliability.

03

Background check

Each strategy is tested on past market periods before being included in the recommendations, reducing the risk of fitting the model to a single scenario.

Decision-making based on mathematical probability

The purpose of the platform is not to predict the market in an absolute sense, but rather to reduce uncertainty in decision-making. Each model expresses its results as a range of probabilities, not a guarantee, allowing the user to assess the risk before making any decision.

Backtesting results are always shown together with the conditions under which they were achieved, as the same strategy may perform differently in a different market environment.

Tareluntik - Illustration of an analytical interface for reviewing predictive models

Investment management tailored to a changing location

  • Low need for manual intervention

    The system performs routine adjustments automatically, and manual review is only required for decisions with a greater impact on the portfolio.

  • Mobile-friendly notifications

    Key alerts are designed for quick review on the phone, without the need to access the entire dashboard.

  • Time zone independence

    Because the analytics work continuously, a change in location or time zone does not affect the timeliness of recommendations.

How the interface is designed

The dashboard is designed for regular but brief checking - a summary of portfolio status, active recommendations and reasons for them are all visible on one screen.

More detailed analytics are available on demand, but not loaded by default, which reduces data consumption when working over mobile networks while traveling.

Risk control as an integral part of any recommendation

Risk is not considered after the fact, but is included in every modeling step, from strategy selection to individual position size.

Algorithmic protection

When the model detects increased volatility, the system automatically suggests hedging positions that reduce exposure without completely exiting the strategy.

Analysis of variance

Each strategy is evaluated against a range of possible outcomes, not just against an average expected value, allowing for a more realistic risk assessment.

Maximum drawdown control

The models take into account the historical maximum decline in value and include it as a limiting parameter when making recommendations.

Distribution of exposure

Recommendations consider diversification across investment classes so that risk is not concentrated in a single strategy or industry.

Your future, optimized.

Join a group of investors who base investment decisions on data and proven models, not random market signals.