Forward-Deployed AI Engineer: Field & Platform Builder

Afresh; San Francisco (CA); Hybrid; USD 168,912 - 273,368; Full time

Benefits offered by this job

Health & wellness

Equity & 401(k) matching

Home office stipend

Job summary

Afresh, the AI platform for grocery, is hiring a Forward Deployed AI Engineer. You will split time between customer field deployments and building the platform that powers those deployments, embedding with data teams, and delivering production‑grade pipelines and AI systems.

This senior, hands‑on role requires shaping architecture with customers, turning field learnings into reusable components, and owning end‑to‑end delivery across field and platform work.

Qualifications

• 5+ years building production software and data systems.
• Architect’s instinct to design clean solutions from messy data.
• Genuine AI/LLM depth with retrieval/agent tech.
• Data-engineering depth; modern cloud data platform.
• Comfort traveling to customer sites regularly (~10-20%).

Responsibilities

• Scope and architect work with customer data sources and production path.
• Embed with customer data teams; integrate into their cloud and data platform.
• Design and ship LLM- and agent-powered systems on that data.
• Harden platform: knowledge graph, retrieval, and grounding layer.
• Build evals, tracing, and tooling for quality and speed.
• Build reusable interfaces and building blocks for multiple customers.
• Own the flywheel: field learnings feed platform.

Skills

Production software

AI/LLM depth

Data engineering

Customer-facing

Cloud data platforms

Architecture

Ownership mindset

Tools

LangGraph

MLOps

Observability

Databricks

BigQuery

Snowflake

Job description

Afresh, the AI platform for grocery, is hiring a Forward Deployed AI Engineer. You will split time between customer field deployments and building the platform that powers those deployments, embedding with data teams, and delivering production‑grade pipelines and AI systems.

This senior, hands‑on role requires shaping architecture with customers, turning field learnings into reusable components, and owning end‑to‑end delivery across field and platform work.