Agentic Harness Engineer

Mind Detect; Wes-Kaap; On-site; ZAR 900,000 - 1,500,000; Full time

Benefits offered by this job

Equity in the business

Generous leave

Medical aid contribution

Advanced equipment/technology

Advanced equipment/technology

Job summary

A technology recruitment firm based in Cape Town is seeking an Agentic Harness Engineer to design and implement tools for AI agents. This hands-on role involves building frameworks to monitor AI performance, collaborating with engineers, and contributing to architecture discussions. Ideal candidates will have a Bachelor's in Computer Science and over 4 years in software engineering. The position offers a generous remuneration package and promotes a great work-life balance.

Qualifications

• 4+ years experience in software engineering, preferably in full-stack or platform engineering.
• Familiarity with LLM-based systems, AI agents, or autonomous workflows.
• Experience with AI agent frameworks and LLM evaluation frameworks.
• Strong understanding of software engineering best practices, including Agile development, source control, testing frameworks, and CI/CD.
• Experience building testing frameworks, evaluation systems, or developer tooling.
• Familiarity with LLM-based systems, AI agents, or autonomous workflows.
• Genuine enthusiasm for agentic AI as the future of engineering.
• Comfort operating as a sole contributor in a greenfield function — no direct reports.
• Experience with AI agent frameworks.
• LLM evaluation frameworks and prompt testing.
• Knowledge of distributed systems and orchestration frameworks.
• Experience with Kubernetes and cloud infrastructure.
• Familiarity with observability tools for AI systems.
• Experience building simulation environments or synthetic test datasets.
• Equity in the business.
• Generous leave/solid work-life balance.
• Great remuneration package.
• Medical aid contribution.
• Plenty of perks.
• An open, international and inclusive culture.
• Advanced equipment/technology.

Responsibilities

• Design and build agent harness frameworks to test and monitor AI agents.
• Develop evaluation pipelines and benchmarking systems for agent performance.
• Diagnose and troubleshoot agent failures and system issues.
• Collaborate with AI/ML engineers, product teams, and platform engineers to ensure agent systems are observable, debuggable, and production-ready.
• Contribute to the engineering team through PRs, architecture discussions, and ceremonies.
• Build tooling and infrastructure to support agent orchestration, simulation environments, and automated testing of agent workflows.
• Participate in architecture design, code reviews, performance evaluation, and internal tool development for agent-based systems.
• Diagnose and troubleshoot agent failures, hallucinations, orchestration errors, and system integration issues.
• Continuously improve agent reliability, observability, and performance by developing robust evaluation harnesses and feedback loops.

Skills

Software engineering best practices

Building testing frameworks

Genuine enthusiasm for agentic AI

Comfort operating as sole contributor

Distributed systems

Kubernetes & cloud

Observability tools

Greenfield solo contributor

Education

Bachelor’s degree in Computer Science, AI, Software Engineering

Tools

Kubernetes

Cloud infrastructure

Observability tools for AI systems

Job description

A unique opportunity for an Agentic Harness Engineer in Cape Town to design and implement the toolkit and infrastructure that enables AI agents to operate within our elite payment orchestration platform’s systems so that the productivity of their workforce can increase exponentially.

The Agentic Harness Engineer is a hands‑on, greenfield tinkerer that will build connectors to internal tooling, provisioning secure execution environments, and ensuring agents adhere to engineering standards before any human review. This is a fast-moving, autonomous contributor role requiring a pragmatic implementer with a genuine enthusiasm for agentic engineering.

Key responsibilities

• Design and build agent harness frameworks to test, evaluate, and monitor autonomous AI agents in production and simulated environments.
• Build and maintain agent connectors to internal systems (e.g. Notion, Google Drive, codebase), managing access permissions to ensure agents operate with appropriate scope and that integrations do not decay over time.
• Develop evaluation pipelines and benchmarking systems to measure agent performance, reliability, and safety.
• Collaborate with AI/ML engineers, product teams, and platform engineers to ensure agent systems are observable, debuggable, and production‑ready.
• Contribute to the engineering team through PRs, architecture discussions, and ceremonies.
• Build tooling and infrastructure to support agent orchestration, simulation environments, and automated testing of agent workflows.
• Participate in architecture design, code reviews, performance evaluation, and internal tool development for agent‑based systems.
• Diagnose and troubleshoot agent failures, hallucinations, orchestration errors, and system integration issues.
• Continuously improve agent reliability, observability, and performance by developing robust evaluation harnesses and feedback loops.

Requirements

• Bachelor’s degree in Computer Science, AI, Software Engineering, or related fields
• 4+ years experience in software engineering, preferably in full‑stack or platform engineering. Note: ML engineering or LLM‑training experience is not required and not the target profile for this role.
• Strong understanding of software engineering best practices, including Agile development, source control, testing frameworks, and CI/CD
• Experience building testing frameworks, evaluation systems, or developer tooling
• Familiarity with LLM‑based systems, AI agents, or autonomous workflows
• Genuine enthusiasm for agentic AI as the future of engineering — candidates who are ambivalent about AI‑assisted development are not the right fit. Personal tinkering or side projects with AI tooling is a strong signal.
• Comfort operating as a sole contributor in a greenfield function — this role has no direct reports and no existing team or playbook to inherit. The successful candidate builds the function from scratch, reporting directly to the Engineering Lead.
• Experience with AI agent frameworks
• Experience with LLM evaluation frameworks and prompt testing
• Knowledge of distributed systems and orchestration frameworks
• Experience with Kubernetes and cloud infrastructure
• Familiarity with observability tools for AI systems
• Experience building simulation environments or synthetic test datasets
• Equity in the business
• Generous leave/solid work‑life balance
• Great remuneration package
• Medical aid contribution
• Plenty of perks
• An open, international and inclusive culture
• Advanced equipment/technology

--

This position is open to people already eligible for work in South Africa

--

About us

We're a dedicated recruiter bringing together the brightest talent with organisations creating cutting‑edge technology to change the world for the better.

We partner with technology providers at the forefront of meaningful innovation. And we’re here for talented individuals who are passionate about using their skills to drive positive change.

Mind Detect provides exceptional recruitment services to businesses who are leading the way in Data, Machine Learning and AI‑driven technologies throughout Europe, the US and Asia.