SparkTrade

Define Your Mandate, We'll Build the Model.

Build a custom equity model from your CIO mandate. Define the mandate. SparkTrade trains, validates, and delivers a machine-learning equity model with a 25-year backtest, current holdings, and an investment-committee-ready research packet.

Start from
Section A · Feature Controls

Technical

each choice ≈ 10 underlying features

Price-action character across short, medium, and long horizons.

Short term
Medium term
Long term

Value

each choice ≈ 10 underlying features

Cheap vs expensive across multiples and yield.

Short term
Long term

Growth

each choice ≈ 10 underlying features

Speed and quality of revenue, earnings, and cash-flow growth.

Short term
Long term

Portfolio construction

Forecast horizon used during training and selection.
Determines how long and short counts relate.
55
11
Log scale · $200M to $3T+
Selected range$2B — $3T+
Eligible equity universe used in training and selection.
How positions are sized within the portfolio.
Drives rebalance frequency, holdings delivery, and subscription price.

Investment policy constraints

Tax-aware construction
Built for RIA investment teams

SparkTrade is for firms that already have an investment philosophy and want a faster way to convert that philosophy into testable, repeatable, model-driven portfolios.

Your firm controls suitability, allocation, trading, and client implementation. SparkTrade supplies the trained model, the validation package, and current holdings.

How SparkTrade is different

vs. no-code trading builders

Tools like Composer convert rules into trades — you bring the strategy. SparkTrade trains a machine-learning model from a mandate. The strategy is the deliverable, not the interface.

vs. quant research platforms

Platforms like QuantConnect give analysts tools and assume they bring the strategy. SparkTrade delivers the trained model and a governed research package, ready for committee review.

vs. direct indexing / rebalancers

Canvas, Vise, and Tamarac personalize and implement portfolios. SparkTrade designs the active model those systems can implement — we complement the stack, not replace it.

What happens next

  1. 01
    Define mandate

    You configure the mandate in the Studio — features, horizon, universe, constraints.

  2. 02
    Train model

    We train a machine-learning equity model across 25+ years of point-in-time data.

  3. 03
    Validate

    Purged K-fold validation with realistic costs, slippage, and survivorship-bias controls.

  4. 04
    Deliver

    Model card, tear sheet, holdings file, and a 30-minute CIO review call.

  5. 05
    Maintain

    Ongoing holdings on your cadence, rebalance notes, model monitoring, and updates.