Staff Machine Learning Engineer, Zeta Global

• 10+ years of experience in machine learning, with 2+ years focused on LLMs, NLP, or generative AI
• Deep understanding of modern LLM stacks and orchestration tools such as LangChain, LangSmith, LlamaIndex, or similar
• Hands-on experience deploying models through Amazon Bedrock, OpenAI API, or comparable inference platforms
• Strong software engineering skills in Python, and familiarity with distributed systems and cloud services (AWS preferred)
• Experience with vector databases (e.g., Pinecone, Redis, PostgreSQL) and RAG pipelines
• Strong grasp of prompt design, evaluation metrics, and fine-tuning techniques
• Ability to collaborate across cross-functional teams and lead architecture-level decisions

Desirable

• We are seeking a Staff Machine Learning Engineer to lead the development of cutting-edge AI applications using LangSmith, Amazon Bedrock, and OpenAI’s foundation models
• As a senior technical leader, you will architect, deploy, and scale generative AI workflows that power intelligent applications across our products
• You’ll collaborate cross-functionally with research, product, and engineering teams to build robust, production-grade systems that push the boundaries of what large language models (LLMs) can do in real-world environments
• Design and implement end-to-end generative AI systems using LangSmith and OpenAI-compatible tools
• Integrate and evaluate foundation models (e.g. , GPT-4, Claude) via OpenAI APIs and Amazon Bedrock
• Develop reusable components and LLM pipelines, including function calling, RAG, agents, and tool augmentation
• Lead efforts in prompt engineering, evaluation, and performance tuning using tools like LangSmith tracing and model analytics
• Define architecture and best practices for scalable, maintainable LLM-based applications
• Mentor junior engineers and guide the team on ML design patterns, evaluations, and deployment strategies
• Drive technical decision-making for LLM inference, vector search integration, latency optimization, and privacy safeguards