AI Engineer
Kobie Marketing4 months ago
Boston, MA, USA +11 moreMid Level
Responsibilities
- Build agent harnesses in Python using LangChain and LangGraph, including tool calling, structured outputs, retries, streaming, and memory.
- Package agent harnesses for the Amazon AgentCore Runtime with appropriate contexts, tools, skills, and subagents.
- Develop agent tools and skills using API integrations, SQL queries against Snowflake, and Snowflake-backed knowledge retrieval with Pydantic validation.
- Build evaluation harnesses using golden datasets, LLM-as-judge approaches, regression suites, and AgentCore Evaluations, and integrate them into CI.
- Implement tool-execution guardrails covering authorization scoping, input and output validation, PII, prompt injection, and hallucination mitigation.
- Deploy, monitor, troubleshoot, and iterate on production agent systems through Amazon AgentCore.
- Partner with data engineers on Snowflake retrieval patterns using Cortex Analyst and Cortex Search Services.
- Contribute to internal engineering patterns as the AI stack evolves.
Requirements
- At least 3 years of professional Python experience building and operating production services.
- At least 1 year of hands-on experience using LLMs in production, including prompt/context engineering, tool or function calling, structured outputs, and RAG.
- Working knowledge of LangChain, LangGraph, AgentCore Strands, CrewAI, Semantic Kernel, or a comparable framework.
- Experience with LLM observability tools such as Amazon CloudWatch, LangSmith, Langfuse, MLflow, or OpenTelemetry.
- Experience designing evaluation frameworks involving MLflow, DeepEval, LLM-as-judge, or multi-turn regression.
- Fluency with Git, Docker, and modern API frameworks.
- Strong written communication and judgment about production readiness.
- Hands-on Amazon Bedrock or AgentCore development experience is strongly preferred.
- Experience with Snowflake, Snowpark, or Snowflake Cortex is strongly preferred.
- Fluency writing and reading SQL and understanding semantic models is strongly preferred.
- Familiarity with multi-agent patterns such as supervisors, routers, subagents, handoffs, reflection, and human-in-the-loop workflows is strongly preferred.
- Experience in loyalty, MarTech, AdTech, or a comparable data-rich B2B domain is strongly preferred.
Benefits
- Remote-first work environment with flexibility to work from anywhere, headquartered in St. Petersburg, Florida.
- Flexible time off.
- Nine company-wide holidays.
- Benefits supporting professional development and personal well-being.
- Periodic travel may be required for business objectives, team collaboration, customer engagements, training, and company events.
Categories
About Kobie Marketing
Kobie Marketing builds loyalty program technology and services for large consumer brands across retail, travel, financial services, telecom, and more. Its offering combines strategy consulting, analytics, and the Kobie Alchemy Loyalty Cloud platform to design, operate, and measure customer rewards and engagement programs. Founded in 1990 and headquartered in St. Petersburg, Florida, the privately held company supports enterprise clients with remote-first teams in the U.S. and India.