# AI Quant Infrastructure

#### Layered capability model

UNIKO’s quant infrastructure is best described as six coordinated layers that form a closed operating loop:

1\.        Quantitative Research Layer

2\.        Execution and Liquidity Support Layer

3\.        AI Optimisation Layer

4\.        Risk Management Layer

5\.        Data Intelligence Layer

6\.        Blockchain and Web3 Infrastructure Layer

Together, these layers aim to create continuous improvement:

·      Data → Research → Execution → Monitoring → Optimisation → Safer execution

#### AI’s role in UNIKO

AI is framed as an enhancement to adaptability, not a replacement for disciplined structure.

AI is used to:

·      Improve recognition and prioritisation of opportunities

·      Adapt parameters based on changing volatility and liquidity states

·      Detect execution regime shifts and reduce false positives

·      Support multi-strategy coordination under unified risk governance

#### Models and parameters (unspecified vs illustrative)

Some technical parameters may be implementation-specific and not publicly fixed. UNIKO can document these as:

·      Unspecified (production): final model architectures, inference latency targets, exact feature sets

·      Illustrative only (examples): ranges and typical engineering targets for transparency

Illustrative examples (non-binding):

<table data-header-hidden><thead><tr><th valign="bottom"></th><th valign="bottom"></th><th valign="bottom"></th></tr></thead><tbody><tr><td valign="bottom">Component</td><td valign="bottom">Production value</td><td valign="bottom">Illustrative example range</td></tr><tr><td valign="bottom">Inference latency target</td><td valign="bottom">Unspecified</td><td valign="bottom">10–200 ms (depending on venue and stack)</td></tr><tr><td valign="bottom">Feature window length</td><td valign="bottom">Unspecified</td><td valign="bottom">1s–30m multi-timescale</td></tr><tr><td valign="bottom">Slippage guard thresholds</td><td valign="bottom">Unspecified</td><td valign="bottom">0.10%–0.80% by tier/liquidity</td></tr><tr><td valign="bottom">Volatility regime bands</td><td valign="bottom">Unspecified</td><td valign="bottom">Low / Medium / High / Extreme</td></tr></tbody></table>

#### Execution engine design principles

A production-grade execution engine typically prioritises:

·      Deterministic behaviour under stress

·      Clear guardrails on slippage and exposure

·      Resilient retry and failover logic

·      Comprehensive logging and observability

#### Monitoring and observability

UNIKO treats observability as a core discipline. Monitoring typically includes:

·      Data health: feed staleness, missing fields, abnormal spikes

·      Execution health: rejects, delays, partial fills, repeated failures

·      Fill quality: slippage, adverse selection, spread capture

·      Risk metrics: inventory skew, exposure concentration, drawdown

·      System metrics: uptime, queue depth, latency distributions

#### Fail-safes and graceful degradation

UNIKO’s fail-safe approach can be documented as:

1\.        Throttle: reduce execution size and frequency when anomalies occur

2\.        Reroute: move to alternate routes/venues when a path degrades

3\.        Pause: temporarily halt a pair or venue when state is unsafe

4\.        Kill-switch: cancel open orders and stop new execution if safety thresholds trigger

5\.        Recover: resume only when health checks pass and stable state is restored


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://uniko.gitbook.io/uniko-whitepaper/ai-quant-infrastructure.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.