Harvest Group is looking to overlay large language models on top of their data warehouse to deliver an AI-driven data experience, enabling secure, accurate, conversational access to data for business stakeholders.
Requirements
- 5+ years in a data architecture role designing analytic data models and platform components (cloud preferred)
- 5+ years in data engineering across ingestion, transformation, modeling, and performance optimization
- 1+ years building LLM‑powered applications or services (e.g., ChatGPT, Azure OpenAI, or comparable) including retrieval, function/tool use, or agents
- Strong proficiency with Snowflake (data modeling, performance, governance, role‑based security)
- Proficient in SQL and a programming language such as Python for data and LLM orchestration
- Experience with Snowflake AI capabilities (e.g., vector search, UDFs/UDTFs, external functions) and/or Snowflake Intelligence/Cortex concepts
- Hands‑on use of LLM frameworks and evaluation tooling; familiarity with prompt engineering and function/tool calling
Responsibilities
- Review and evolve current data structures—models, schemas, and naming/metadata conventions—in our data warehouse to make them more LLM/AI‑ready (clear semantics, consistent business definitions, performant patterns, and governed access).
- Design, implement, and maintain the semantic and metadata layers that sit between LLMs and our data warehouse to enable governed, conversational access to data.
- Build and operationalize unstructured contextual data pipelines (e.g., documents, PDFs, presentations, images, logs, tickets), including ingestion, parsing/OCR, chunking, metadata extraction, embedding generation, indexing, and integration into the AI/RAG platform.
- Architect secure RAG (retrieval‑augmented generation) patterns, including query planning, grounding, and context assembly from our data warehouse and related systems.
- Stand up rapid proofs of concept and experiments; iterate quickly to de‑risk approaches, then harden winning patterns into enterprise‑ready solutions.
- Partner with business end‑users to understand questions and decisions, then translate them into schemas, policies, and conversational intents the system can reliably support.
- Create clear technical documentation and enablement materials for internal users and support teams.
Other
- This role is based in Rogers (AR) or Cincinnati (OH). If you are applying but do not currently reside in one of these markets, please note that relocation will not be covered by Harvest Group.
- Demonstrated ability to communicate with both technical and non‑technical partners and to lead through influence
- Proven track record of shipping systems from prototype to production with attention to reliability, cost, and maintainability
- Working in an Agile Scrum framework with rapid, iterative delivery
- CPG retail industry experience and familiarity with retailer data ecosystems
