Mid/Senior Agentic Systems Developer

Autonomous AI Agent Supervision and Software Delivery

This is not a conventional developer role. The successful candidate will not simply write code manually or passively accept AI-generated outputs. They will act as a technical supervisor for a suite of AI agents, ensuring that agent-generated work is accurate, secure, maintainable, commercially useful, and aligned with real software requirements.

The role suits a developer who understands modern software engineering and can also manage the behaviour of AI systems as they plan, reason, generate, test, and refine outputs across changing models and technical environments.

About Coral Code

Our work focuses on building practical technology that improves productivity, reduces operational friction, and creates measurable commercial value.

The Role

You will review agent-generated code, guide agent workflows, assess technical decisions, improve prompts and system instructions, evaluate model performance, manage edge cases, and intervene when autonomous outputs require human engineering judgment.

You will also help shape the way Coral Code uses AI agents internally, including how agents are structured, monitored, evaluated, and improved over time.

Key Responsibilities

AI Agent Supervision and Orchestration

• Supervise autonomous AI agents working on software development, documentation, testing, analysis, and delivery tasks
• Break larger technical objectives into clear work packages that can be handled by specialised agents
• Review agent reasoning, outputs, code changes, documentation, and proposed technical decisions
• Detect when an agent is producing plausible but incorrect, incomplete, unsafe, or commercially unsuitable work
• Create and refine agent instructions, workflows, guardrails, escalation rules, and review checkpoints
• Coordinate multiple agents working across related tasks, ensuring outputs remain consistent and aligned
• Monitor agent performance across changing models, system behaviour, context windows, cost profiles, and output quality
• Maintain human oversight over critical technical decisions, production risks, security-sensitive work, and client-facing deliverables

Software Development

• Design, build, review, and maintain web applications, SaaS modules, workflow systems, dashboards, APIs, and integrations
• Convert business requirements into practical software architecture and implementation plans
• Review and improve code generated by AI agents before it enters production workflows
• Build reliable front-end and back-end functionality using modern software development practices
• Write clean, maintainable, secure, and well-structured code
• Debug complex issues across application logic, data flow, integrations, and user workflows
• Participate in code reviews, technical planning, release preparation, and quality control

AI Evaluation and Quality Control

• Create evaluation criteria for agent-generated code, plans, documentation, and technical recommendations
• Test AI-generated outputs against functional requirements, edge cases, security expectations, and maintainability standards
• Compare outputs from different model types and select the best approach for the task
• Identify failure patterns such as hallucinated logic, weak assumptions, incomplete testing, over-engineered solutions, or hidden security risks
• Improve feedback loops so agents become more useful, consistent, and aligned with delivery standards
• Maintain a practical balance between speed, cost, reliability, and engineering quality

Governance, Security, and Risk Management

• Ensure AI agents operate within defined access, data, security, and approval boundaries
• Apply least-privilege principles when agents interact with codebases, systems, documentation, APIs, or business data
• Ensure sensitive information is handled carefully in AI-assisted workflows
• Define when human approval is required before agent-generated work progresses
• Help maintain traceability over agent decisions, generated outputs, changes, and review history

Required Experience

• Mid to senior level software development experience
• Strong understanding of software architecture, code quality, testing, debugging, and maintainability
• Practical experience working with AI systems beyond basic code completion
• Ability to evaluate outputs from different AI models and understand where each model is strong or weak
• Experience working with structured prompts, system instructions, workflows, model comparison, and iterative output improvement
• Ability to supervise autonomous or semi-autonomous AI workflows
• Strong understanding of APIs, databases, application logic, front-end development, back-end development, and system integrations
• Ability to work independently and make sound technical decisions
• Strong written communication skills for technical documentation, agent instructions, and delivery notes
• Good judgment around security, privacy, data handling, and production risk

Technical Skills

• TypeScript, JavaScript, React, Node.js, Python, or comparable modern development stacks
• API design and integration
• SaaS platforms and workflow systems
• Relational databases and structured data models
• Cloud environments, preferably Azure or comparable platforms
• CI/CD, version control, testing, and deployment processes
• AI-assisted development pipelines
• Multi-model reasoning and model selection
• Retrieval-augmented generation, structured outputs, tool calling, agent memory, or agent workflow design
• Observability, logging, evaluation frameworks, and quality metrics for AI-generated work

What Makes This Role Different

This role is for someone who can manage that shift.

You should be able to answer questions such as:

• Which parts of this task should be done by an AI agent and which require human judgment?
• Which model or agent setup is best suited to this task?
• What could go wrong if this agent output is accepted too quickly?
• How should this workflow be structured so the agent produces better outputs next time?
• How do we measure whether the agent is improving productivity without reducing software quality?
• How do we keep the system safe, traceable, and commercially useful?

The Person We Are Looking For

• A practical developer with strong technical judgment
• Comfortable supervising AI-generated work without being impressed by surface-level fluency
• Able to detect weak logic, missing requirements, poor assumptions, and hidden implementation risks
• Interested in the future of software delivery, not just today’s coding workflows
• Disciplined enough to use AI for speed while maintaining engineering standards
• Commercially aware and focused on building software that solves real business problems
• Clear, structured, and calm when managing complex technical work

What You Will Work On

• SaaS product modules
• Client workspace and portal systems
• Workflow automation platforms
• Compliance and regulated-industry software
• AI-assisted business process automation
• CRM and operational systems
• Dashboards, reporting, and analytics
• Internal AI agent workflows
• Documentation, testing, and delivery acceleration systems

Why Join Coral Code

You will help shape how Coral Code builds software faster, improves productivity, reduces delivery friction, and maintains quality as AI capabilities continue to change.

Apply now

Share