SparkTrade

Quant as a Service Capabilities

License a production strategy, stream daily predictions, build models with our tools, or have our quants build one to your mandate. One quant team runs all four.

Licensed strategies

Multiple machine-learning production models, ready to license.

Long-only, long/short and short-only models with regime-aware, low-volatility, high-volatility and all-weather variants. They provide a distinct view of today's quant-driven market. Firms use them for model portfolios, separately managed accounts, platform strategies and research.

  • Defined characteristics and rebalance schedules
  • Daily scores or holdings, delivered by file or feed
  • Four signal sets on the Snowflake Marketplace

Delivered as

Holdings filesAPISnowflake

For: Advisors, fund managers, platforms, data teams

Ask about licensing
Profiles
Long-onlyLong/shortShort-only
Variants
Regime-awareLow-volatilityHigh-volatilityAll-weather
Used for
Model portfoliosSeparately managed accountsPlatform strategiesResearch

Prediction streams

Daily signal feeds on about 3,500 U.S. stocks: long and short AI scores, prediction-market event research and news-driven sentiment.

AI Signals

Every trading night, each stock gets a long score and a short score from 0 to 100. Momentum models look for moves that are building. Catch-up models look for stocks that have lagged and may recover. Each score shows the factors behind it and its signal history. Scores are rankings, not probabilities.

  • Long and short scores from 0 to 100
  • Momentum and catch-up models
  • Factor view and signal history for every stock

Delivered as

Platform moduleAPIMCPSnowflakeFiles

For: Platforms, data teams, advisors

Event Pressure

Prediction-market odds, turned into company research.

Event Pressure watches prediction markets, where traders put real money on the outcome of elections, rate decisions, policy changes and economic data. When the odds move or major news breaks, it identifies the stocks and ETFs most exposed and explains the likely direction of pressure, the mechanism, the evidence and the timing.

  • Prediction-market odds mapped to the stocks and ETFs most exposed
  • Direction of pressure on each company or sector, and why
  • Links to the markets and sources behind each finding
  • Confidence and timing for each finding
  • Company research only, with no buy or sell calls

Delivered as

Platform moduleAPIMCP

For: Platforms, data teams

News & Sentiment

News & Sentiment tracks trending news and social conversation, matches stories to the companies involved, and estimates the direction and size of likely impact, how fast each story is spreading and which related companies it may reach.

  • Story-to-ticker mapping
  • Direction and size of likely impact
  • How fast a story is spreading
  • Related companies that may feel second-order effects

Delivered as

Platform moduleAPIMCP

For: Platforms, data teams

Algo creation tools

Tools that turn a trader's history or a client's rules into a working model.

Mockingbird

Mockingbird describes patterns in a trader's closed trades and flags new setups that match them. It uses a read-only brokerage connection or a CSV import. Training starts once the trader has enough closed trades.

  • Read-only brokerage connection or CSV import
  • A summary of the trader's patterns, such as holding periods and setups
  • A personal model once there are enough closed trades
  • Alerts when a new setup matches

Delivered as

Platform module

For: Platforms and their clients

  1. 1

    History

    Connect a read-only brokerage account or import a CSV of closed trades.

  2. 2

    Style

    See the trader's patterns in holding periods and setups.

  3. 3

    Model

    Train a personal model once there are enough closed trades.

  4. 4

    Alerts

    Get an alert when a new setup matches.

Quant Design Studio

Self-serve model building for clients of our platform partners. Clients can start from a recommended preset or choose their own factors and constraints. Studio then trains and tests the model. Institutions use Quant Design Studio on our site to define a Mandate to Model build.

  • Recommended presets to start from
  • Set each factor to raise or lower a score
  • Testing on periods the model never saw

Delivered as

Platform module

For: Platforms and their clients

Discuss platform access

Mandate to Model

A custom model, built and run by our quants.

Define your mandate in Quant Design Studio. SparkTrade builds and tests the model, provides the research package, then hosts the model and delivers holdings for your SMA or model portfolio, or through a data feed. Model files are not transferred.

  • 25-year backtest, tear sheet and model card
  • Current holdings as of the build date
  • A review with our quants
  • Ongoing holdings with model monitoring

Delivered as

Hosted model with holdings files

For: Advisors, fund managers, prop firms and institutional clients of our platform partners

Open Quant Design Studio
  1. 1

    Define

    Set the factors, horizon, universe, position counts and constraints in Quant Design Studio.

  2. 2

    Train

    We train a machine-learning model on 25 years of point-in-time data.

  3. 3

    Test

    We test the model on periods it never trained on, with realistic costs and slippage.

  4. 4

    Deliver

    You get the model card, tear sheet, holdings and a review with our quants.

  5. 5

    Run

    We host the model and deliver holdings on your schedule, with monitoring and updates.

Not sure where to start?

Tell us who you serve and what you want to offer. We'll tell you which capability fits.

Book a call