AI Engineer

Role overview Build AI-powered applications and agents for tasks such as support, analyst assistance, document processing, and automation. You will take systems from implementation through evaluation and monitoring, with attention to security risks such as prompt injection and misuse. The role combines software engineering with practical measurement of quality, safety, latency, and cost.

Responsibilities - Develop model-powered features and agents using model APIs and frameworks, including tool use and structured outputs. - Design retrieval-augmented generation pipelines, covering document chunking, embeddings, retrieval, and re-ranking. - Build tools and servers that give agents appropriately limited and validated access to other systems. - Create evaluation datasets and automated checks for quality, safety, and cost, and run them in continuous integration. - Add safeguards such as input and output validation, permission-aware retrieval, and human approval for higher-risk actions. - Instrument systems with tracing and monitoring, investigate failures, and compare models and prompts against real tasks.

Requirements - Around 2–3 years of software engineering experience, including at least one year building LLM- or ML-powered features. - Strong Python skills; TypeScript is useful. - Hands-on experience with language-model APIs and prompt design. - Experience building retrieval-augmented generation systems with a vector store. - Understanding of prompt injection, data leakage, and hallucination risks and how to reduce them. - Familiarity with evaluation practices for AI applications and comfort with Git, testing, and continuous integration and delivery.