We Build Agentic Research Systems

Institutional Standards. Skin in the Game.

We built and ran production platforms at BlackRock, Barclays Global Investors, and Morgan Stanley.

Every model and signal runs on our own capital first.

What We Do

Proprietary Research & Signals

Agentic systems for idea generation, backtesting, and live signal production — deployed on our own capital.

  • Cross-sectional ML ranking and time-series strategies under unified governance
  • Strict walk-forward evaluation with point-in-time data integrity
  • Ensemble methods across linear and nonlinear learners with quality gates
  • In-house stack: TimescaleDB, Airflow, AWS Bedrock, SageMaker, Python/Java/C++

Selective Private Mandates

We partner with a small number of aligned trading cohorts, prop shops, and family offices each year.

  • Full research-to-production delivery: data ingestion through live deployment
  • Multi-sleeve fund construction with institutional cost modelling
  • Allocator-grade reporting, audit trails, and mandate compliance
  • Production architecture with governance, promotion workflows, and event logging

Platform

We don't build one-off strategies. We operate a systematic investing platform — a composable architecture where any alpha source becomes a reusable sleeve, passes through the same fund builder, overlay stack, cost engine, and reporting pipeline.

Research Engine

  • Structured feature engineering: momentum, quality, distress, macro regime, relative value
  • Walk-forward training with strict point-in-time correctness — no lookahead, ever
  • Multi-model ensembles (Ridge, PLS, LightGBM, XGBoost) with IC-weighted blending
  • Statistical validation: PSR, deflated Sharpe, bootstrap CIs, HAC-corrected t-stats

Fund Construction

  • Sleeve abstraction: any strategy wraps into a composable unit with standardised outputs
  • Inverse-volatility-correlation allocation across sleeves
  • Overlay stack: momentum, volatility regime, drawdown control — multiplicative scaling
  • Short-book realism: borrow cost, recall risk, margin, and prime-broker economics

Institutional Plumbing

  • Three-level NAV accounting: research gross, implementation net, investor net after fees
  • Full audit trail with immutable run manifests and deterministic replay
  • Reconciliation, cost waterfalls, capacity reporting, and mandate compliance
  • Cloud-native on AWS with live execution via Interactive Brokers prime brokerage APIs

Working With Us

We primarily manage our own books. Selectively, we take on private mandates where there is genuine alignment in strategy, timeframe, and incentives.

Good-fit mandates typically involve systematic equity strategies, multi-factor research platforms, or production infrastructure for small to mid-sized asset managers and prop shops with a 6–18 month horizon.

Principals

15+ years building and operating systematic research platforms across BlackRock, Barclays Global Investors, and Morgan Stanley. Three principals. Own capital deployed alongside every mandate.

Locations

London · New York · Barcelona · Malta · Budapest · Lisbon · Geneva

We embed directly with partner teams across time zones.

If there is alignment, let's talk.

info@hecatusresearch.com