pubX builds ML-driven pricing and monetization tools for digital publishers, with a tech stack built for real-time decision-making: Kafka for event streaming, FastAPI + Node.js for service layer, PyTorch + scikit-learn for modeling, and AWS infrastructure (SageMaker, Bedrock, EKS, Lambda). The project mix—real-time pricing decisions, MLOps infrastructure, agentic AI backends, and lakehouse data patterns—reveals a company solving the manual and static pricing problems that cost publishers significant revenue, now moving toward autonomous optimization systems.
pubX is a London-based AI startup (founded 2020) building monetization technology for digital publishers. The core product is a Dynamic Price Floor Tool that applies machine learning to set prices individually per ad auction, replacing manual and rule-based approaches. The company operates across real-time bidding infrastructure, ML pipeline orchestration, and revenue analytics, serving mid-market and enterprise publishers. The engineering-heavy hiring profile (5 senior-level roles in past 30 days, split across engineering, data, and product) indicates active scaling of core ML and infrastructure capabilities.
Python, TypeScript, FastAPI, Kafka, PyTorch, scikit-learn, AWS (SageMaker, Bedrock, EKS, Lambda, RDS, DynamoDB), Terraform, MLflow, and Weights & Biases for ML operations and infrastructure.
Real-time pricing decisions for ad auctions, production ML pipelines, MLOps infrastructure, agentic AI backends, and lakehouse data models for publisher monetization and revenue optimization.
Other companies in the same industry, closest in size