What are the responsibilities and job description for the AI Agent Builder position at Harrison Clarke?
AI Agent Engineer / AI Agent Builder (Startup)
We’re working with a well-funded, early-stage AI company building next-generation agentic systems for enterprise workflows.
This isn’t about simple chatbots or wrappers around APIs - this team is focused on designing production-grade AI agents that can reason, take action, and operate reliably in complex, real-world environments.
They’re looking for engineers who enjoy working at the intersection of LLMs, systems design, and product, and who want to push beyond prompt engineering into true agent architecture.
What You’ll Work On
- Designing and building multi-step, agentic workflows powered by LLMs
- Developing tool-calling systems that integrate with real enterprise software
- Building robust RAG pipelines, memory systems, and context management layers
- Creating evaluation frameworks to measure and improve agent performance
- Shipping production-ready AI systems with strong reliability, observability, and guardrails
What They’re Looking For
- Experience building LLM-powered systems in production (not just prototypes)
- Strong understanding of agentic architectures, tool use, and orchestration
- Hands-on experience with RAG, vector databases, and retrieval systems
- Ability to design scalable backend systems (Python, distributed systems, APIs)
- Familiarity with frameworks like LangChain, LangGraph, LlamaIndex, or similar
- A mindset focused on evaluation, reliability, and real-world performance
Bonus Points
- Experience with multi-agent systems or complex orchestration patterns
- Exposure to LLM evaluation, post-training, or fine-tuning workflows
- Built systems involving memory, planning, or long-running agent tasks
- Background in early-stage startups or building products from scratch
Why This Role
- Work on cutting-edge AI systems beyond typical GenAI applications
- Join a small, high-calibre team with strong technical leadership
- Own systems end-to-end, from design to production
- Meaningful impact on real enterprise use cases, not demos