Senior Applied AI Engineer – Prompting & Evaluation Job Stellenbosch South Africa,Software Development

"At Klipboard we've introduced a flexible hybrid work policy, where employees spend three days in the office and two days working from home. This approach promotes a balanced work environment that combines office collaboration with the comfort and convenience of remote work."

Klipboard provides specialist software, services and support to deliver fully integrated trading and business management solutions to companies in the distributive trade – wherever they are in the world. With a unique depth of knowledge and experience in ERP/SaaS solutions, Klipboard has a wide range of clients includes wholesalers, distributors, merchants and retailers from small traders to multinational enterprises. Klipboard has offices in the UK, Ireland, The Netherlands, South Africa, Kenya and North America.

Our mission is simple: to design and deliver high performance, integrated ERP solutions that enable our distributive trade customers to source effectively, stock efficiently, sell profitably and service competitively.

Klipboard is a global, growing business that embraces AI and emerging technologies to enhance customer outcomes, collaboration, and continuous improvement. We’re looking for people who are curious about or fluid with AI, open to change, and excited to learn how technology can improve the way we work and help our customers which is always supported by strong human insight and communication.

This is a high-impact senior engineering role for someone who wants to shape how a large engineering organisation uses AI in real products and real development work. You will set patterns, create evaluation standards, coach engineers and help Klipboard turn AI from individual experimentation into repeatable, scalable engineering practice.

The role is two jobs in one: the craft of getting reliable, measurable results from large language models through prompt engineering, context design and evaluation; and making everyone around you better h hundreds of engineers across the UK and South Africa, the difference between an organisation where a few people prompt well and one where everyone does is enormous. Your success is measured less by what you personally produce and more by how quickly the engineers around you improve.

Key Responsibilities:

Role Accountabilities

Practise the craft at a high level – designing, building and refining prompts, context strategies and agentic workflows for AI features in our products.

Grow the engineers around you – coaching, pairing, workshops and building shared assets that spread good practice across hundreds of engineers.

Set standards and keep the organisation honest – defining what good looks like for prompting, evaluation and context engineering, and maintaining those standards as tools change.

Key Activities and Contributions

Design, build and refine prompts, context strategies and agentic workflows for AI features in Klipboard products, where a confidently wrong price, part match or stock answer is worse than no answer.

Build evaluation into everything – test prompts against real business cases at scale, measure quality honestly, and treat “it looked fine when I tried it“ as the start of the work, not the end.

Work across models and providers, choosing pragmatically on quality, cost and latency rather than loyalty, and keeping up as models, tools and providers change quickly.

Get deep into the domain detail with product teams and subject matter experts, because a prompt is only as good as the understanding of the business problem behind it.

Coach engineers across teams to get dramatically more from AI coding tools and LLMs, through pairing, reviews, workshops and patient one-to-one help.

Build and maintain shared prompt patterns, reusable templates, evaluation harnesses and internal guides, so good practice spreads as assets rather than folklore.

Run an internal community of practice where engineers share what is working, what failed and why, and make it somewhere people actually want to show up.

Help define what good looks like for prompting, context engineering and evaluation across R&D, and keep those standards current as tools and models change.

Share measurable results with engineering…