Prompt Engineer

The company has integrated AI into several customer-facing and internal workflows — but output quality is inconsistent, prompts have been written by whoever needed them, and there's no systematic approach to evaluating what works or why. The difference between a prompt that performs reliably and one that doesn't is costing real time and real money.

We're looking for a Prompt Engineer to take ownership of the AI layer — auditing what exists, building a structured approach to prompt design and testing, and working with the product and engineering teams to ship AI features that actually perform in production.

What you'll work on

• Auditing existing prompts across all active AI integrations — identifying inconsistencies, failure modes, and optimization opportunities
• Designing and iterating prompts for production use cases — classification, generation, extraction, summarization, and conversational flows
• Building a systematic evaluation framework — defining success criteria, creating test sets, and measuring prompt performance quantitatively
• Integrating prompts into product features via OpenAI API and Claude/Anthropic API, working closely with the engineering team
• Implementing advanced techniques — chain-of-thought, RAG, tool use, and multi-agent orchestration using LangChain or similar frameworks
• Evaluating new model releases and capabilities against existing production use cases as the landscape evolves

Requirements

• 2+ years of hands-on prompt engineering for production AI applications — not just experimentation or side projects
• Deep familiarity with OpenAI and Anthropic model families — their strengths, limitations, and behavioral differences across versions
• Experience building evaluation frameworks and test sets for LLM output quality
• Comfortable in Python for scripting, testing pipelines, and API integration