Generative AI Engineer

Dataiku•United States, New York, NY

23h•$160,000 - $240,000•Hybrid

About The Position

As a Generative AI Engineer on the ED&A team, you will build the agentic AI systems that change how Dataiku runs internally. The role is hands-on and end-to-end: you’ll work close to the business, turn real problems into working software, and see your solutions through from first conversation to production. This position can be based in our New York office or remotely within the Eastern Time Zone.

Requirements

Must have strong Python skills (including familiarity with typical data science and AI engineering libraries).
Must have hands-on experience building agentic AI systems, multi-agent orchestration, tool chaining, autonomous decision-making, and production deployment of AI agents.
Experience with Go-to-Market motions, nomenclature, and technology, including Salesforce, HubSpot, 6sense, Gong, Outreach, and related tools.
Experience with modern agent orchestration frameworks (LangGraph, CrewAI, Claude Agent SDK, OpenAI Agents SDK, or similar); familiarity with LangChain is still relevant but not sufficient on its own.
Understanding of RAG architectures (vector databases, embedding strategies, agentic RAG, GraphRAG) and when to apply each approach.
Familiarity with MCP (Model Context Protocol) for agent-to-tool integration, or demonstrated ability to quickly adopt new integration standards.
Experience with structured outputs, function/tool calling, and prompt engineering across multiple LLM providers.
Web development fundamentals (HTML, CSS, JavaScript); experience with Vue.js and Node.js preferred.
Exposure to AI evaluation practices, building evals, monitoring model/agent performance in production, and iterating based on metrics.
Comfort with AI-assisted development tools (GitHub Copilot, Cursor, Claude Code, or similar).

Nice To Haves

Familiarity with Dataiku a bonus.

Responsibilities

Design end-to-end AI solutions on Dataiku's platform, leveraging Dataiku Agent Hub, Prompt Studio, LLM Mesh, and Knowledge Banks (Vector Stores), or Python-based frameworks where needed.
Build and orchestrate multi-agent systems using Dataiku's Visual Agents (simple and structured), as well as code-based frameworks (LangGraph, CrewAI, Claude Agent SDK, OpenAI Agents SDK) as appropriate.
Integrate and optimize LLM APIs across providers (OpenAI, Anthropic, Google Gemini, AWS Bedrock, Azure, open-source models via Dataiku's LLM Mesh), applying model routing strategies to balance cost, latency, and quality.
Implement Retrieval-Augmented Generation (RAG) pipelines, including agentic RAG and GraphRAG, using Dataiku's Knowledge Banks with reranking, dynamic filtering, and document extraction capabilities.
Work exclusively with the Marketing organisation, partnering across functions such as Demand Generation, Content Marketing, Product Marketing, Field Marketing, Marketing Operations, Brand, and Communications.
Engage marketing stakeholders to gather business requirements, then go further: identify the underlying user or team pain points those requirements represent, and design solutions that address both the stated need and the deeper problem.
Own projects end-to-end, from requirements intake and solution design through to build, deployment, and handover.
Develop autonomous and semi-autonomous AI agents, using Dataiku's Agent Builder, custom Python-based architectures (LangGraph, CrewAI, Claude Agent SDK, etc.), or a combination of both. Exercise judgment on when to leverage platform capabilities and when to build custom solutions.
Design and build Agent Tools beyond documented examples, including custom API integrations, data retrieval modules, decisioning logic, and automated workflows, pushing past out-of-the-box patterns to deliver solutions tailored to specific business problems.
Build, publish, and consume MCP (Model Context Protocol) servers to enable agent-to-tool integration across systems, including designing custom MCP servers where needed.
Develop evaluation and monitoring approaches for agent systems, combining Dataiku's built-in capabilities with custom instrumentation to measure reliability, accuracy, cost, and business impact in production.
Design and maintain evaluation frameworks (evals) for LLM-based systems, measuring accuracy, latency, cost, and reliability in production.
Adhere to data governance, security, and regulatory compliance requirements (EU AI Act awareness, responsible AI practices) for all AI solutions.
Leverage Dataiku's Cost Guard and Quality Guard features to manage LLM spend, enforce usage policies, and maintain output quality standards.
Work closely with analytics and data engineering teams to maintain metadata on reference datasets for LLM consumption.
Create front-end user interfaces for AI applications using HTML, CSS, and JavaScript, within Dataiku's webapps framework, Dataiku Answers for chat-based interfaces, or standalone applications built with Vue.js and Node.js.
Collaborate on UX design, ensuring internal stakeholders find AI solutions intuitive and responsive.
Provide product feedback to the development team to improve the platform.
Stay current with the rapidly evolving AI engineering landscape, agent frameworks, model capabilities, evaluation practices, governance requirements, and tools like MCP and A2A protocols.