About
Data Platforms Built for Decisions
We pull data from disparate operational systems and external sources and consolidate it into high-performance analytical warehouses. From there we build dashboards, predictive models, and integrations that make the information actionable — whether that means a real-time wagering system, a multi-year e-commerce operating platform, or domain-specific ML driving nine-figure capital allocation.
Key Projects
Select Engagements
E-Commerce Operations Platform — Perelel
Multi-year engagement with Perelel Health beginning at launch in 2020, continuing through exponential revenue growth. We built the company's core data infrastructure, integrating every meaningful source across operations, inventory, fulfillment, sales, customer service, advertising, customer research, and financials into a unified analytical platform. The platform integrates with external partners for marketing attribution and inventory management, built to scale and continuing to grow with the business.
Bloodstock Acquisition & Management
Developed, own, and operate a data platform and ML pipeline for thoroughbred horse evaluation and acquisition. The system consolidates pedigree, genetic, racing, veterinary, sales, and biomechanical data into a unified model used to assess future potential and value. Predictive models drive purchasing and breeding decisions for broodmares, stallion shares, and racehorses.
Parimutuel Wagering
Developed, own, and operate real-time predictive systems for parimutuel horse racing markets. The platform ingests live odds, track conditions, weather, and historical performance data, running ensemble models that produce continuously updated probability distributions for race outcomes. Identifies positive expected-value opportunities in real time and integrates directly with wagering infrastructure for execution.
Core Capabilities
Technical Areas of Practice
Multi-Source Data Integration
Ingesting and unifying data from ERPs, CRMs, third-party APIs, flat files, and event streams into a single normalized data model — handling schema reconciliation, deduplication, and entity resolution across sources.
High-Performance Warehouses
Deploying and optimizing columnar analytical databases including StarRocks, Snowflake, ClickHouse, and PostgreSQL-based solutions.
Multi-Modal Data Handling
Tabular, graph, spatial (GIS), medical (DICOM), genomic, photos, videos, audio, and more — integrated and operated on within unified pipelines.
Data Normalization & Modeling
Dimensional and entity-centric data models with incremental transformation frameworks (DBT, SQLMesh). Semantic layers and metrics definitions for consistent, governed access to trusted data.
Predictive Modeling & ML
Production machine learning using XGBoost, PyTorch, Bayesian hierarchical models, and distributional regression — from experimentation through deployment.
Service Deployment & Integration
ML models and data products deployed as services that push predictions, scores, and enriched data back into operational systems, dashboards, and third-party platforms.
Infrastructure & DevOps
On-premise and cloud Kubernetes clusters (Talos Linux, EKS), GPU compute for model training, GitOps with Flux CD, and hybrid cloud architectures spanning AWS and bare-metal environments.