AI Systems Engineer (Agentic Development)

About the role

Soar, a global fintech startup focused on financing and investment, is looking for an AI Systems Engineer to design, manage, and optimise AI agents that generate code, workflows, and business logic. You will ensure that AI‑generated outputs are correct, scalable, and aligned with the company’s real‑estate and mortgage business goals.

Key responsibilities

• Design and manage agent‑based development workflows.
• Oversee AI‑generated code: validate, test, refactor, and productionise.
• Build systems to evaluate agent output quality and reliability.
• Fine‑tune prompts, tools, and pipelines to improve outcomes.
• Integrate AI agents into the software development lifecycle (SDLC).
• Translate mortgage and real‑estate business problems into AI‑driven solutions.
• Establish best practices for multi‑agent collaboration and orchestration.
• Monitor performance, cost, and accuracy of AI systems.
• Collaborate with product and business teams to identify automation opportunities.

Required profile

• Hands‑on experience with LLM‑based tools and workflows (e.g., Anthropic Claude or similar).
• Experience building or managing AI‑assisted or agent‑driven systems.
• Strong understanding of modern SDLC with AI acceleration.
• Experience with Google Cloud Platform or comparable cloud environments.
• Ability to review and validate generated code across multiple languages.
• Experience with prompt engineering, evaluation loops, and model tuning.
• System‑thinking mindset: able to design end‑to‑end workflows, not just write functions.

Required skills

• LLM tools (Anthropic Claude or similar)
• AI‑assisted/agent‑driven system development
• Google Cloud Platform (GCP)
• Prompt engineering
• Model tuning and evaluation loops
• Multi‑language code review
• Vector databases and RAG pipelines
• Multi‑agent orchestration frameworks

What we offer

• Build an AI‑first company from the ground up.
• Work on high‑impact real‑estate and mortgage problems.
• Fast, experimental, innovation‑driven environment.
• Shape engineering practices, not just deliver features.