Senior Software Engineer

About The Role

About Soar and how we build

Soar is building an AI-native home-financing platform for Saudi Arabia, and the AI-driven engineering system that builds it. Our product spans origination, credit decisioning, loan management, collections and investor products, all under SAMA regulation.

We build differently. Our engineers don't hand-write code; they specify, direct and review AI agents that do. If that sounds like where engineering is going, you'll fit here.

How we work :

Agents write the code. Engineers own the spec, the context, the tests and the review. DevOps doesn't hand-manage servers either.

Context engineering is the core skill. Agents work from a shared knowledge graph of business rules, architecture principles and money-movement rules that we keep current.

We build our own agents. Requirements, DevOps, PM and self-testing agents run on our MCP/A2A framework to automate the SDLC end to end.

Full stack by default. AI closes skill gaps, so engineers work across backend, frontend and infrastructure, with cross-domain reviewers.

Regulated means rigorous. Money-movement logic gets human review, every time. KSA data residency shapes which models and tools we use and where they run.

What every role here needs: you already use AI coding agents (Claude Code, Cursor, Codex or similar) as your main way of working, not as autocomplete. You can show us real workflows you've built, where they broke, and what you changed.

Role Summary:We rely heavily on AI throughout engineering and are looking for a Senior Software Engineer who treats AI agents as a core part of how they build software. You will own complex features from design through production, direct agents effectively, and apply strong engineering judgment to everything you ship. You should be equally comfortable accelerating implementation with AI and investigating difficult problems yourself.Key Responsibilities:

• Design, build, and maintain production features and services, taking ownership of correctness, performance, security, and maintainability.
• Use AI agents across planning, coding, refactoring, debugging, test generation, and documentation to accelerate end-to-end delivery.
• Provide agents with clear requirements, architecture context, constraints, and acceptance criteria; break complex work into verifiable steps.
• Inspect and refine AI-generated code. Detect incorrect assumptions, insecure patterns, unnecessary complexity, and missing edge cases.
• Build automated unit, integration, and end-to-end tests; validate actual user journeys and failure paths before declaring work complete.
• Create reusable workflows and tool integrations that help agents access the right context and perform tasks within appropriate permissions.
• Own changes through review, deployment, observability, and production troubleshooting; investigate root causes and prevent repeat failures.
• Collaborate with Product, QA, and other engineers; mentor teammates and document decisions and validation evidence in pull requests.
  
  

Requirements

Requirements:

• Typically 5+ years building and shipping production software, or equivalent demonstrated engineering depth and ownership.
• Strong proficiency in at least one production programming language and experience with application design, APIs, databases, version control, and automated testing.
• Substantial hands-on experience delivering software with AI coding tools or agents, supported by concrete examples of shipped work.
• Ability to reason about, debug, and explain code independently, including code generated by AI.
• Practical understanding of context management, tool use, agent limitations, and the trade-offs between model capability, speed, and cost.
• Strong ownership and collaboration, including careful handling of credentials, sensitive data, and permissions in AI workflows.
  
  

Nice to have:Experience with agent orchestration, MCP or API integrations, retrieval systems, evaluation suites, cloud platforms, or financial systems.What success looks like:You ship reliable features with clear validation evidence, catch defects in generated code, improve delivery speed, and create reusable practices that raise the effectiveness of the wider team.

---