AI Engineer

Responsibilities

• Design and build AI-powered features across our B2B portal, internal tooling, and merchant-facing products — including LLM integrations, AI agents, and intelligent automations
• Translate ambiguous business problems into well-scoped AI solutions, from prompt engineering and RAG pipelines to full agentic workflows
• Build, evaluate, and iterate on AI systems using a rigorous experiment-driven approach — tracking quality, latency, and cost tradeoffs
• Collaborate closely with product, operations, and engineering teams to identify high-leverage AI opportunities and deliver them end-to-end
• Develop internal AI tooling and skill frameworks that empower non-technical teams to leverage AI in their daily workflows
• Integrate with third-party AI APIs (Anthropic, OpenAI, etc.) and MCP-based tooling while maintaining security and reliability standards
• Maintain observability over deployed AI systems — monitoring for regressions, prompt drift, and model performance degradation
• Work independently in a remote environment with a strong sense of ownership and ability to ship with minimal oversight

Requirements

• 3+ years of software engineering experience, with at least 1–2 years focused on building production AI or ML systems
• Hands-on experience with LLM APIs (Anthropic Claude, OpenAI GPT, etc.) and prompt engineering best practices
• Strong programming skills in Python and/or TypeScript/JavaScript; comfortable building both backend services and lightweight frontend interfaces
• Experience building RAG pipelines, embedding workflows, or agentic systems using frameworks like LangChain, LlamaIndex, or similar
• Familiarity with vector databases (Pinecone, Weaviate, pgvector, etc.) and semantic search patterns
• Experience working cross-functionally with non-technical stakeholders to scope and deliver AI projects
• Proven ability to evaluate AI output quality and build evals/testing frameworks for LLM-based systems
• Logistics, supply chain, or B2B SaaS experience is a strong plus
• Experience with MCP (Model Context Protocol), AI agent orchestration, or multi-step tool-use workflows is a bonus