For serious investors. This website is not an offer or solicitation to invest.

Systematic fund management

Systematic by design. Intelligent by construction.

Corellian Global trades systematically across all major asset classes: equities, fixed income, currencies, commodities and their derivatives. Our strategies are built on AI-driven signal research and overseen by people who have spent decades running trading businesses for AAA investment banks.

Signal ensemble Illustrative
Individual signals → combined exposure Not performance data
Markets All major asset classes
Research AI-assisted signal discovery
Oversight Human sign-off on every model change
Status Live strategy testing
Approach

Machines find the patterns. Experience decides what deserves capital.

We treat AI as a research instrument, not an oracle. Every signal has to survive the same questions a senior risk committee would ask of a human trader.

Built by traders, not just technologists

Our team has run trading businesses for AAA investment banks. We build models the way we managed risk on the desk, with a clear view of how strategies fail in live markets.

Evidence before conviction

Signals are tested out of sample, across regimes and under realistic costs before they reach a portfolio. A strong backtest alone earns nothing.

Many small edges

The portfolio combines a diversified ensemble of signals, so no single model, market or idea dominates the outcome.

Explainable by design

We can tell an investor why the strategy holds a position. Models we cannot explain do not trade.

Investment process

Six stages from raw data to an executed order.

Each stage has a named owner. Automated where precision and speed matter, human where judgement does.

01

Data

Market, macro and alternative data are cleaned, time-stamped and checked for look-ahead bias before any research begins.

Automated
02

Signal research

AI agents generate and stress-test candidate signals; researchers reject anything that does not hold up economically.

AI + researcher
03

Trade decisions

The portfolio manager decides each trade using the AI's insight. Position sizes follow fixed rules set against a volatility budget, with correlation and concentration limits.

Portfolio manager
04

Risk review

Exposures, drawdown limits and liquidity are checked before trading and monitored continuously after.

Human sign-off
05

Execution

Approved trades are executed under fixed rules to limit market impact, with slippage measured against arrival price.

Rules-based
06

Review and evolution

Live results are compared with research expectations. Models that drift are retrained, re-weighted or retired.

Head of Evolution
Risk framework

Risk is set first. Returns are what is left inside it.

Human judgement and machine intelligence, working together around the clock.

Our aim is to deliver market-beating returns within managed drawdown limits. The AI monitors positions and risk 24 hours a day, seven days a week, and our team acts on what it finds.

That partnership between human and machine is what makes the strategy complete and resilient.

  • Portfolio volatility targetDefined annualised budget
  • Drawdown controlsStaged de-risking
  • Single-market concentrationCapped by risk share
  • LeverageGross and net limits
  • LiquidityDays-to-exit tested daily
  • Model kill-switchAutomatic on breach
People

Built by people who have run trading floors.

Over half a century of combined experience in global markets, investment banking and risk.

Founder & Chief Investment Officer

Graeme King

25 years running global trading businesses for AAA investment banks.

Head of Systematic Trading

Alex Vanderlip

30 years in investment banking and consultancy with UBS and KPMG.

Partner, Head of Evolution

Alan Thomas

Former Head of Structured FX Sales at Merrill Lynch and Head of Wealth Management at Credit Suisse.

Chief Risk Officer

Chris Ashton

Former Global Head of FX Trading at Barclays.

Fund status

Where we are today.

We publish no performance figures until the strategy has a live, verifiable record.

  1. Completed

    Research and model build

    Signal library and systematic trading agent developed.

  2. Now

    Live strategy testing Current

    Three-month testing phase of the systematic trading agent, running to 31 December 2026.

  3. Next

    Fund structuring and registration

    Vehicle, administrator and regulatory approvals, targeted for completion by the end of 2026.

  4. Then

    Seed capital and launch

    Founding investor allocations, with reduced fees on the first $10 million raised.

Indicative terms
Investor eligibility
Professional only
Minimum subscription
To be confirmed
Management / performance fee
To be confirmed
Liquidity
To be confirmed
Founding investor terms
Reduced fees on first $10m
Investor enquiry

Talk to us about founding allocations.

Request the information pack or a call with the investment team. We respond to professional investors within two business days.

This field is for validation purposes and should be left unchanged.
Professional investor confirmation

Important information

This website is for information only and is directed solely at professional and institutional investors. It does not constitute an offer to sell, or a solicitation of an offer to buy, any interest in any fund or investment product. Any such offer will be made only through the fund's offering documents.

The fund described is in development and is not yet authorised or registered. Strategy descriptions reflect current intentions and may change. Graphics marked illustrative are not based on, and do not represent, actual or simulated performance.

Trading in financial instruments involves substantial risk, including the loss of all capital invested. Systematic and AI-driven strategies carry model risk and may behave differently in live markets than in research. Past performance is not a reliable indicator of future results.

[Legal entity name], registered in [jurisdiction], company no. [number]. Registered office: [address].

[Regulatory status: authorised and regulated by (regulator), ref. (number), or not yet authorised.]