SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Latitude is hiring a Prompt Engineer to design, optimize, and maintain prompts for production GenAI and LLM applications. This role focuses on advanced prompt engineering techniques and LLM optimization to improve application quality and performance.
Key Responsibilities:
- Design, optimize, test, and maintain prompts for production GenAI/LLM applications
- Develop and implement advanced prompting strategies including zero-shot, few-shot, system/role prompting, structured outputs, and multi-step prompting
- Optimize instructions, retrieved content, conversation history, examples, and metadata to improve LLM response quality
- Build and execute evaluation frameworks using golden datasets, scoring rubrics, regression tests, and A/B testing
- Troubleshoot prompt failures and implement quality metrics (accuracy, relevance, groundedness, faithfulness, hallucination rate)
- Design and test safety measures including prompt injection detection, jailbreak prevention, and data leakage mitigation
- Design multi-turn conversations with clarification strategies and fallback handling
- Implement RAG and grounding strategies using embeddings, semantic search, and citation-based approaches
- Integrate with LLM APIs and platforms, managing tokens, latency, and costs
- Maintain prompt versioning, documentation, CI/CD pipelines, and production monitoring
- Translate business requirements into measurable prompt performance improvements
Requirements:
- 2+ years of hands-on experience designing, optimizing, testing, and maintaining prompts for production GenAI/LLM applications
- Expertise in zero-shot, few-shot, system/role prompting, structured outputs, prompt templates, multi-step prompting, dynamic prompts, and instruction hierarchy
- Strong understanding of LLM behavior, hallucinations, tokenization, context windows, model parameters, latency, and cost
- Hands-on experience with OpenAI/Azure OpenAI, Claude, Amazon Bedrock, or Gemini
- Ability to build evaluation frameworks using golden datasets, scoring rubrics, regression tests, and A/B testing
- Understanding of embeddings, chunking, semantic search, retrieval quality, and context relevance
- Experience testing prompt injection, jailbreaks, data leakage, and harmful outputs
- Familiarity with function calling, tool use, MCP, multi-agent workflows, LangChain, LangGraph, Semantic Kernel, or Prompt Flow
- Strong Python skills with REST APIs/SDKs, JSON/structured outputs, streaming, token management, and model configuration
- Experience with prompt versioning, Git, documentation, CI/CD, regression testing, production monitoring, and LLM traces
- Ability to translate business requirements into prompts and measurable quality improvements
- Notice period: Immediate to 15 days
