LLM Application Engineer

LLM Application EngineerLocation: Remote, hybrid, or onsiteEmployment type: Full-timeAbout the roleWe’re hiring an LLM Application Engineer to build AI-powered product features using large language models.You’ll work across model integration, prompting, evaluation, retrieval, tool use, and production application development. The role sits between applied AI and software engineering: turning model capabilities into features that users can rely on repeatedly.You’ll work with product, engineering, and design teams to take LLM-based features from prototype through production.What you’ll doDesign, build, test, and maintain applications powered by large language models.Integrate models from providers such as OpenAI, Anthropic, Hugging Face, and open-source ecosystems.Build prompting, retrieval, tool-calling, and orchestration workflows.Evaluate model behaviour using automated tests, datasets, and production feedback.Improve output quality, latency, reliability, and inference cost.Build APIs and services that expose AI capabilities to product applications.Experiment with model selection, prompting strategies, retrieval techniques, and fine-tuning where appropriate.Monitor deployed AI features and investigate failures or unexpected model behaviour.Work with product and engineering teams to define requirements and technical trade-offs.Document system behaviour, evaluation methods, and implementation decisions.What we’re looking forExperience building applications using LLMs or other machine-learning models.Strong programming experience with Python.Experience integrating LLM APIs or open-source models into production software.Understanding of prompting, embeddings, retrieval, and model evaluation.Experience with frameworks such as PyTorch, Hugging Face, or equivalent ML tooling.Experience building or consuming APIs.Ability to test non-deterministic systems and identify failure modes.Understanding of production concerns including latency, reliability, security, and cost.Ability to explain model behaviour and engineering trade-offs to technical and non-technical colleagues.A degree in Computer Science, Machine Learning, AI, or a related subject can be useful, but equivalent practical experience is also accepted.Nice to haveExperience with retrieval-augmented generation.Experience building agentic or multi-step LLM workflows.Experience with tool calling and structured model outputs.Experience fine-tuning or adapting language models.Experience building evaluation datasets and automated LLM evaluation pipelines.Experience deploying open-source models.Experience with vector databases and semantic search.Experience building AI services using APIs or microservice architectures.Experience working with multimodal models.What a typical week may includeYou might spend your time:Prototyping an LLM-powered product feature.Comparing model outputs across prompts, models, or retrieval strategies.Building evaluation cases for a new AI workflow.Investigating why a production interaction failed or produced an incorrect result.Improving retrieval, tool use, or orchestration for a multi-step workflow.Working with backend engineers to expose an AI capability through an API.Reviewing production traces to identify opportunities to reduce latency or inference cost.Testing a new model against existing quality and reliability requirements.How we workLLM features are treated as production systems rather than demonstrations.We test model behaviour against defined use cases, measure failures in production, and use those results to decide what to improve next.Engineers are expected to own features from experimentation through deployment and monitoring, while working with product, design, backend, and client engineers throughout the process.Hiring approachYou do not need to match every item in this description to apply. We assess candidates based on the skills required to do the job and the experience they can bring to the team.We welcome applications regardless of gender, race, ethnicity, religion, disability, sexual orientation, age, or background.