AI Prompt Engineer VA
Designs, tests and maintains the prompts, system instructions and evaluation sets behind your AI assistants and automations, so outputs are accurate, on brand and consistent. Turns a flaky chatbot into a dependable tool your team and customers can rely on.
- Top 1% vetted talent
- No lock-in contracts
- Replacement guarantee
Get an Instant Quote for AI Prompt Engineer VA Hire
Hourly rates · No lock-in contracts
Entry Level
$9 / hr
Intermediate ★
$12 / hr
Senior
$15 / hr
Responsibilities
- Write and iterate system prompts, few-shot examples and structured output schemas for customer-facing assistants and internal AI workflows.
- Build evaluation datasets and scoring rubrics, then run regression tests in tools such as Promptfoo, LangSmith or Braintrust before any prompt change goes live.
- Design multi-step prompt chains and agent workflows that call tools, retrieve documents and hand off to humans when confidence is low.
- Tune retrieval settings, chunking and context windows so RAG assistants cite the right source and stay within cost budgets.
- Measure hallucination, refusal and format-failure rates and drive them down with prompt, model or pipeline changes.
- Compare models from OpenAI, Anthropic, Google and open-weight providers on quality, latency and cost for each use case, and recommend the fit.
- Write brand voice and safety guidelines into prompts and verify tone against real transcripts.
- Maintain a versioned prompt library with change notes so every production prompt can be traced and rolled back.
- Train support, sales and operations staff to use AI tools well and collect their failure cases for the next iteration.
Software & Tools
ChatGPT
Claude
OpenAI API
Anthropic API
LangChain
Promptfoo
Python
Requirements
- 5+ years in a technical, analytical or content role with at least 2 years of hands-on prompt engineering for production LLM applications.
- Deep familiarity with the GPT, Claude and Gemini model families, their APIs, function calling and structured output features.
- Working Python or JavaScript for building test harnesses and small pipelines with LangChain, LlamaIndex or the raw SDKs.
- Experience with prompt evaluation tooling such as Promptfoo, LangSmith or OpenAI Evals, and version control with Git.
- Understanding of embeddings, vector databases and retrieval-augmented generation.
- Excellent written English with a precise, plain style, since the prompts themselves are the deliverable.
- Reliable high-speed internet with a power backup, and comfortable overlapping the client's working day for review sessions.
Outsource Your Next Hire With Confidence
Scale your business faster with our vetted assistants from the Philippines. Start your free one week trial today and experience the difference.