Senior AI Engineer

Role Overview

You will own an AI application module or agentic workflow end to end, including the parts most teams skip: evaluation, guardrails, cost and reliability. This is production AI engineering rather than prototyping. You will work directly with the client product owners who define how a feature should behave, and we keep teams small and senior, so there is little distance between your work and the users who depend on it. If you have shipped a GenAI feature that real users rely on, and know how much harder that is than the demo, this role is built for you.

Technical Skills

• Languages & Frameworks: Python, REST API development, and async patterns for handling concurrent model calls.
• LLM Systems: Provider APIs such as OpenAI and Anthropic. Prompt engineering beyond trial and error, structured output handling, and function or tool calling.
• Agent Orchestration: LangChain, LlamaIndex or an equivalent framework, plus multi-step and multi-agent workflow design.
• Retrieval: RAG architecture end to end — chunking strategy, embedding models, vector databases such as Pinecone, Weaviate, pgvector or FAISS, and retrieval quality tuning.
• Evaluation & Safety: Building evaluation harnesses, defining quality metrics for generative output, regression testing prompts, and guardrail and content filtering patterns.
• Production & Deployment: Cost and latency optimisation, caching, streaming responses, fallback strategy and model versioning. Observability with LangSmith, Langfuse or equivalent. Containers or serverless, CI/CD, and deploying inference workloads on at least one major cloud.

Good to Have

Not required, but a strong plus if your experience already includes any of the following.

• Experience building evaluation harnesses or guardrail frameworks from scratch
• Fine-tuning and adapter methods such as LoRA
• Exposure to AI security and adversarial testing, including prompt injection defence
• Prior mentoring experience, formal or informal