Data Scientist 3

Data Scientist 3As AI continues to evolve, there is a need to quickly learn new technology and build AI Agents and use them to create alerts & automate workflows. Most of the company's structured data resides in Snowflake. We are looking for an expert who has experience building AI Agents in Dataiku & Snowflake and using Dataiku recipes & Python for workflow automation.We are using AI to identify disruptions in the supply chain and take corrective action to avoid backorders on parts. This position will hit the ground running in creating AI Agents to solve this use case. We also need to upskill the current team in the latest Dataiku Agentic AI technology. This position will share their knowledge and teach this new technology to team members. The position will focus on strengthening our capabilities in agent architecture, orchestration, security, integrations, evaluation, and operationalization, while establishing best practices that can be reused across multiple AI programs and use cases.Understand the objective - Write code in Snowflake & Dataiku to build AI Agents - Test output and update codeDaily standup meetingsWork on a project with 3 other people (includes project leader)OfficeYears of experience: 5-7years - Degree requirement: Bachelors in AI, Data Science, Computer Science - Do you accept internships as job experience: Yes - Are there past or additional job titles or roles that would provide comparable background to this role: NoBuilt agents in Dataiku & Snowflake Cortex (critical) - structured data - Workflow automation in Dataiku recipes & Python (write programming logic) - Experience with LangGraph, Langchain, CrewAIMulti-agent architecture and orchestration patterns o Agent collaboration, planning, delegation, and workflow orchestration o Reusable agent frameworks and reference architectures - Security, access control, and governance o Role-based access control (RBAC), personas, authorization models, and guardrails o Auditability, and compliance considerations - Enterprise integrations and data access patterns o Integration with Microsoft Teams, SharePoint, email, PDFs, and enterprise data sources o Event-driven and batch processing architectures o API and workflow integration best practices - Agent evaluation, testing, and observability o Offline and online evaluation methodologies o Groundedness, relevance, hallucination detection, and task-success measurement o Automated regression testing, monitoring, observability, and performance tracking - Prompt engineering and agent design patterns o Prompt optimization techniques, reusable prompts, skills, tools, and agent design frameworks - Human-in-the-loop workflows o Approval processes for high-impact business decisions o Escalation paths, feedback loops, and continuous improvement mechanisms - Dataiku application development and user experience o Dataiku WebApps, user experience design, and production-ready front-end development o Best practices for enterprise adoption and usability - Deployment, operationalization, and scalability o Deployment automation, monitoring, and governance - Cost and performance optimization o LLM selection and deployment strategies o Latency, token consumption, and operational cost management o Performance tuning and optimization best practicesClientP experience in multi-agent system to coordinate communication across agents (Agent to Agent)Team playerGood communication skills