Principal AI Engineer
Zeta Global6 hours ago
Berlin, GermanyStaff+
Responsibilities
- Define AI/ML architecture, shared capabilities, technical roadmaps, interfaces, and build-versus-buy decisions across the portfolio.
- Lead predictive ML and optimization work for ranking, recommendation, targeting, bidding, pacing, classification, and personalization.
- Design and build reusable agentic frameworks involving orchestration, retrieval, tool integration, state, memory, and multi-agent coordination.
- Own production implementations across data pipelines, model training, deployment, inference, MLOps, and LLMOps.
- Establish reproducibility, versioning, automated testing, observability, safe rollouts, and rollback practices.
- Define offline benchmarks and online experiments for model quality, agent performance, latency, cost, and advertiser outcomes.
- Set standards for reliability, drift detection, data quality, incident response, privacy-aware access, auditability, and protection against prompt injection and data leakage.
- Partner across engineering, data science, product, and business teams; lead design and code reviews, mentor senior engineers, and resolve architectural decisions.
Requirements
- 12+ years of relevant software engineering and applied ML experience, or equivalent demonstrated expertise, with substantial ownership of production ML systems and technical leadership across multiple teams.
- Deep expertise in predictive ML for ranking, recommendation, classification, or optimization, including statistics, experimentation, model evaluation, and the full model lifecycle.
- Hands-on experience delivering production LLM applications and agentic workflows involving tool use, retrieval, orchestration, evaluation, guardrails, and operational monitoring.
- Experience architecting distributed, high-throughput, low-latency systems and balancing model quality, infrastructure cost, and operational complexity.
- Strong Python skills and proficiency in Java, Go, or a comparable backend language.
- Experience with PyTorch or TensorFlow, SQL, and large-scale data or streaming systems such as Spark and Kafka.
- Experience with cloud infrastructure, containerization, model serving, production observability, and relational, NoSQL, caching, and retrieval systems.
- Ability to influence without direct authority, communicate architectural decisions, mentor experienced engineers, and translate business objectives into measurable technical outcomes.
- Preferred qualifications include AdTech experience, recommendation or marketplace systems, causal inference, uplift modeling, contextual bandits, constrained optimization, online learning, privacy-aware identity and audience systems, email advertising, and reusable ML or agent platforms.
Tech Stack
Categories
About Zeta Global
Zeta Global builds an AI-powered marketing cloud that unifies customer data, analytics, and omnichannel activation for enterprise brands and agencies. Its Zeta Marketing Platform combines martech and adtech—spanning CRM, email, display, and social—to acquire, grow, and retain customers on a subscription and media-buying model. Founded in 2007 and headquartered in New York City, Zeta is a public company trading on the NYSE under the ticker ZETA.