Applied AI Engineer - Prompting & Evaluation in Lamesley

Overview

Hands-on role building AI features into established C# .NET codebases. Design prompts and context, integrate LLMs, and deliver production-grade capabilities with strong evaluation and maintainability.

Responsibilities

• Design and build prompts, context strategies and LLM integrations for product features, in domains where a confidently wrong price, part match or stock answer is worse than no answer.
• Work primarily in C# .NET, integrating AI capability into established codebases through clean service boundaries, sensible abstractions and respect for the code that is already there.
• Prototype quickly for real deadlines, harden in weeks, and distinguish between permissible and non-permissible optimizations.
• Build evaluation alongside the feature, test against real business cases, measure quality, and let results inform decisions.
• Handle non-glamorous aspects: error handling, fallbacks when a model misbehaves, latency, token cost, logging and monitoring.
• Collaborate with engineers who own each codebase to fit in with existing patterns and pipelines.
• Keep up with changing models, tools and providers and make pragmatic quality, cost and latency trade-offs.
• Share learnings with engineers through code, examples and conversation.
• Work with product managers, product owners and subject matter experts to understand the business problem properly.

Systems, Tools and Technology

• C# .NET (primary development language)
• Large language model APIs across multiple providers
• AI coding tools: GitHub Copilot, Cursor or equivalents
• Prompt engineering and context design patterns
• Retrieval-augmented generation (RAG), vector search, embeddings (desirable)
• Evaluation frameworks and automated quality pipelines for AI outputs

Technical and Professional Expertise

• Solid production experience with C# .NET, including working in established codebases you did not write, and shipping changes into them safely.
• Hands-on experience building with large language models: prompt design, context engineering and structured outputs, in real work.
• A track record of shipping quickly, taking ideas to working software in weeks rather than quarters.
• Experience testing or evaluating LLM outputs in a structured way, and using results to improve quality.
• Daily fluency with AI coding tools such as GitHub Copilot, Cursor or equivalents.

Core Responsibilities and Contributions

• Deliver AI features end-to-end: from requirement understanding through to shipped, evaluated product capability.
• Maintain production-grade quality: error handling, fallbacks, latency management, logging and monitoring.
• Care about accuracy, safety and data handling; customers rely on the answers our software provides.
• Document work to facilitate future maintenance by other engineers.
• Prototype fast, harden properly, and know where to cut corners and where not to.

Customer Experience

• Ensure AI features deliver accurate, trustworthy answers; avoid confidently wrong responses.
• Understand the business problem behind each feature, not just the technical solution.
• Collaborate with product managers and subject matter experts to ensure AI capability serves customer needs.

Key Outcomes and Activities

• AI capability shipped into at least one established product with evaluation within the first six months.
• Something built has gone from idea to customers in weeks, and held up in production.
• Evaluation results influence decisions, potentially killing approaches not suitable for shipping.
• Engineers around you adopt techniques from your work; teaching is not your primary job.
• You can explain the business problem behind each feature, not just the technical solution.

People, Collaboration & Culture

• Bias to action: prefer smaller, working versions over lengthy planning.
• Honest about quality: measures, shows working software, and does not ship unmaintainable work.
• Respect existing code and the engineers who maintain it.
• Curious about the domain in Klipboard’s customers to inform better prompts.
• Comfortable with change; tools evolve and adaptability is valued.