Platform overview

Live execution

Live monitoring of portfolio performance, side by side with the simulated numbers.

Trade Monitor: realised vs simulated PnL

A full set of live statistics, sliced by strategy, long/short, sector and security.

Intraday performance by long/short with metric selector

Detailed analysis of performance and risk metrics, all updated in real time.

Performance breakdown: realised vs simulated with commission, slippage and delay

The portfolio at a glance: current versus target positions, notionals and other customizable fields.

Portfolio positions with targets, notionals and liquid hours

The order blotter: every fill with broker, algo, average price, status and commission.

Orders blotter with fills, brokers and commissions

Features:

  • Automated front, middle and back office: live execution without building the infrastructure.
  • Live performance continuously compared against the simulation, with slippage and drift analysis.
  • Risk management and position monitoring across strategies and accounts.
  • Review trading activity from anywhere (phone, tablet, desktop, etc).

Backtesting

Every backtest dissected: statistics and equity curves by sector, long/short, instrument type and more.

Backtest performance by sector with statistics and equity curves

Correlations across strategies, to build portfolios that diversify rather than overlap.

Correlation heatmap across strategies

Convexity analysis: strategy returns against the benchmark.

Convexity analysis of strategy returns versus benchmark

From strategy implementation to analysis, in one integrated research environment.

Strategy implementation and analysis in the research lab

Features:

  • Realistic backtesting: costs, slippage and liquidity modelled so results survive live trading.
  • Parallel simulation on the cloud or in calc-farms, on daily/snapshot and minute data.
  • Interface in Python, implementation in C++.
  • Portfolio construction and risk-analysis tools.
  • Proprietary reports to analyze trading strategies:
    • Sensitivity analysis
    • Costs analysis
    • Parameters & overfitting analysis
    • Analysis by long/short, instrument type, sector, etc.

AI strategy lab

The lab at work: the agent forms a hypothesis, edits the code, backtests it with realistic costs and validates it on a holdout it has never seen.

The lab agent iterating: hypothesis, code edit, backtest and holdout validation

Every experiment tracked in a ledger: hypotheses, results and verdicts, accumulated while you sleep.

Experiment ledger with hypotheses, statuses and pass/fail verdicts

Each candidate ends in a research note written by the agent, with its own caveats, ready for human review.

Research note written by the agent with results and caveats

Features:

  • AI agents that design, code, backtest and refine systematic strategies, from idea to tested candidate.
  • Proprietary research methods developed in-house by a team of experts in statistics, finance and machine learning, including patented strategy-evaluation technology (US patent 12,271,950).
  • Every candidate is validated by the same realistic backtesting engine used for live strategies.
  • The IP of every strategy developed with the lab is owned by the client.

Reporting

PnL analysis: realised, live and simulated results compared day by day, with cumulative differences tracked over time. Anomalies can be investigated down to the level of each security.

PnL analysis report: realised vs live vs simulated with cumulative differences

Cost analysis: commissions, slippage and delays measured in basis points, live versus simulated, by instrument type and down to each security. Over time, this allows building increasingly accurate cost estimates for every security.

Cost analysis report: total costs in basis points, live versus simulated

Features:

  • Daily comparison of realised, live and simulated results.
  • Cost attribution in basis points, by instrument type and down to each security.
  • Date-range filters and one-click export on every table.
  • Custom reporting is also available upon request.

Data cleaning

We reconcile your data sources with our proprietary algorithms, or simply provide our own clean data.

Data cleaning: multiple sources reconciled into clean data

Features:

  • Automated error detection and correction, with tools for manual review.
  • Data distributed as analytics via cloud.
  • Supports stocks, ETFs, futures, options, and currencies.

Operations

The entire daily operation runs as scheduled jobs, from data snapshots to compliance copies, each with a clear status.

Scheduler with the daily operation jobs and their statuses

Every process monitored in real time, with live logs for each job.

Scheduler showing running jobs with live logs

Full manual control when needed: re-run, edit or force any step directly from the dashboard.

Manual job controls: re-run, edit and force options

Automated checks across execution, processes, system and real-time data, with every breach flagged immediately.

Automated checks dashboard across execution, processes, system and real-time data

Features:

  • Continuous trade monitoring with automated alerts, so strategies are watched after they go live.
  • Automated reconciliation of cash, positions and trades.
  • Compliance controls and logging built into the trading workflow.
Request a Call