Trireme is a digital-asset trading, market-making and quantitative research firm operating across crypto markets, systematic trading and emerging financial-market structures.

‍

We are expanding our quantitative research capability into prediction markets and event-driven trading.

‍

We are looking for a Quantitative Researcher / Trader who can develop systematic strategies for markets where prices represent the probability of real-world outcomes.

You will research, build and operate quantitative strategies across prediction markets.

‍

The role requires a combination of:

• Probability and statistics.
• Quantitative research.
• Market microstructure.
• Event-driven trading.
• Data engineering.
• News and information processing.
• Risk management.

You should be capable of thinking about markets as probability distributions rather than simply directional price charts.

‍

The objective is to identify situations where implied market probabilities materially differ from the probability of an event occurring, and determine whether that difference represents genuine edge after liquidity, execution costs, settlement mechanics and model uncertainty are considered.

• Research systematic and discretionary-assisted strategies across prediction markets.
• Build models estimating probabilities of real-world events.
• Compare model-derived probabilities against market-implied probabilities.
• Identify and quantify mispricing.
• Develop event-driven signals using structured and unstructured data.
• Build data pipelines for relevant economic, political, financial, sporting or other event data.
• Research cross-market and cross-platform arbitrage opportunities.
• Analyse order books, spreads, liquidity and market depth.
• Model expected value and position sizing.
• Develop execution strategies appropriate for relatively illiquid markets.
• Monitor changing information and update probability estimates dynamically.
• Build systems for rapid response to material news or data releases.
• Research market-maker behaviour and participant biases.
• Analyse historical calibration of prediction markets.
• Conduct rigorous backtesting where historical market data is available.
• Build simulation frameworks where traditional backtesting is not possible.
• Monitor live strategy performance.
• Maintain clear P&L and attribution by strategy and event.
• Develop risk controls around correlated event exposures.
• Monitor platform, counterparty and settlement risk.
• Document model assumptions and sources of uncertainty.
• Work closely with engineers to automate data ingestion, pricing and execution.

Research may include:

• Elections and political events.
• Macroeconomic releases.
• Central-bank decisions.
• Regulatory events.
• Crypto-specific events.
• Token launches and protocol milestones.
• Financial-market outcomes.
• Geopolitical events.
• Sports or other highly liquid event markets.
• Cross-platform probability discrepancies.
• Prediction-market versus traditional-market pricing.
• News sentiment and information extraction.
• Poll aggregation.
• Bayesian updating.
• Forecast combinations and ensemble models.
• Market calibration and participant behavioural biases.

The specific markets traded will depend on liquidity, legality, platform access, risk and demonstrated edge.

Strong candidates should be comfortable with concepts including:

• Conditional probability.
• Bayesian inference.
• Calibration.
• Expected value.
• Kelly-style position sizing.
• Monte Carlo simulation.
• Logistic models.
• Time-series analysis.
• Ensemble forecasting.
• Alternative data.
• NLP and information extraction.
• Correlated outcomes.
• Scenario analysis.

The ability to assign uncertainty to your own probability estimate is as important as producing the estimate itself.

If a market implies a 40% probability of an event occurring, it is not enough to conclude that the contract is attractive because your model predicts 50%.

‍

You should be able to explain:

• Why your estimate is 50%.
• The confidence interval around that estimate.
• Which assumptions drive the discrepancy.
• Whether those assumptions are already reflected in the market.
• How quickly your informational advantage is likely to decay.
• Available liquidity.
• Expected slippage.
• Settlement and platform risk.
• Correlation with the existing portfolio.
• Appropriate position size.
• The conditions that would invalidate the trade.

• Strong background in quantitative research, statistics, mathematics, computer science, economics, finance, physics or another analytical discipline.
• Excellent probability and statistical reasoning.
• Strong Python.
• Experience working with financial or event-driven datasets.
• Ability to build models from incomplete and noisy information.
• Strong understanding of expected value and uncertainty.
• Ability to distinguish genuine informational edge from narrative conviction.
• Strong research discipline.
• Intellectual honesty and willingness to update views as evidence changes.
• Ability to operate independently.

Experience with:

• Prediction markets.
• Polymarket or similar venues.
• Event-driven trading.
• Sports betting or quantitative betting markets.
• Polling models.
• Forecasting competitions.
• Bayesian modelling.
• NLP / LLM-based information extraction.
• News-data pipelines.
• Alternative data.
• Crypto market structure.
• Market making.
• Automated execution.

A successful background in quantitative sports betting can also be highly relevant given the overlap in probability estimation, pricing and market efficiency.

Candidates should be expected to complete a practical research exercise.

‍

An example assessment could involve providing:

• A live or historical prediction market.
• Current market-implied probability.
• A set of relevant data sources.

The candidate would then be asked to:

1. Produce an independent probability estimate.
2. Explain the methodology.
3. Identify the most important variables.
4. Quantify uncertainty.
5. Determine whether a trade exists.
6. Recommend a position size.
7. Define invalidation criteria.
8. Explain the main risks.
9. Describe how the model would update as new information arrives.

We care more about the quality of reasoning and calibration than whether the candidate happens to predict the correct eventual outcome.

Success will be judged on:

• Calibration of probability forecasts.
• Expected value captured.
• Realised versus expected P&L.
• Risk-adjusted returns.
• Maximum drawdown.
• Forecast accuracy.
• Brier score or equivalent calibration measures.
• Quality of research.
• Repeatability of strategies.
• Execution quality.
• Model robustness.
• Speed and quality of information processing.
• Performance after transaction costs.
• Strategy scalability.
• Risk-limit adherence.

Prediction-market positions can contain significant hidden risks.

‍

The role will be expected to explicitly manage:

• Correlated event exposures.
• Binary payout risk.
• Liquidity gaps.
• Event-resolution ambiguity.
• Platform and counterparty risk.
• Settlement risk.
• Data-source errors.
• Model overconfidence.
• News latency.
• Concentrated political or geopolitical exposure.
• Sudden probability repricing.

No strategy should move into meaningful production capital without a documented hypothesis, defined risk limits and evidence of positive expected value.

A strong researcher in this role should eventually be capable of building a system where:

‍

Data → probability estimate → uncertainty → market comparison → expected value → position sizing → execution → monitoring → probability update → exit/settlement

‍

is increasingly systematic, measurable and automated.

Competitive base or contractor fee with performance-linked upside.

‍

For the right candidate, compensation may include a meaningful variable component linked to validated strategy performance and realised trading P&L, subject to agreed risk limits and loss-adjusted performance.