AI Engineers (George)

Job description

AI Engineer (Pro-Code)
Junior | Mid | Senior
Badger SA | George, Western Cape

We build AI that gives customers a soft landing and gives teams their time back.
At Badger, AI isn’t a side project. It’s how we scale engagement, service, and internal execution.

We’re hiring a pro-code AI Engineers to architect and ship production-grade agents and conversational systems - RAG-based, deeply integrated with our CRM, portals, and messaging channels.
If you can turn powerful models into simple, reliable experiences (fast, accurate, on-brand, and governed), you’ll fit here.

What you’ll own

Build production AI systems

• Architect, build, and run conversational agents and automated workflows that deliver measurable efficiency (time saved, faster resolution, lower cost-to-serve).
• Turn model capability into product-grade behaviour: fast, accurate, resilient, and easy to operate.

Integrate and orchestrate the stack

• Connect AI systems to messaging channels (e.g., WhatsApp, Microsoft) and orchestrate across our CRM and self-service portals.
• Design clean integration patterns (APIs, events, connectors) so data flows reliably end-to-end.

Operate like an engineer, not an experimenter

• Use telemetry, tracing, and DevOps practices to monitor performance, troubleshoot issues, and iterate quickly.
• Improve speed, stability, and quality continuously - without breaking trust.

Partner with the business (without losing the plot)

• Work with business leaders to shape AI features and prioritise what matters
• Explain complex behaviour and data insights in plain language. Make tradeoffs explicit.

Guardrails, governance, and security

• Build with privacy and ethics as defaults: POPIA/GDPR.
• Treat governance as a design constraint, not paperwork.

Multiply the team

• Maintain clean version control, documentation, and runbooks.
• Host deep dives and mentorship to raise the technical bar across the team.

What we’re looking for (must-haves)

• Strong Python and real experience shipping with frameworks/APIs such as Azure OpenAI, OpenAI API, LangChain, Microsoft agent frameworks, or similar.
• Hands-on experience with Azure AI Foundry (deployment, experimentation, evaluation; fine-tuning where relevant).
• Proven work building RAG / knowledge-grounded systems (retrieval strategy, chunking, embeddings, reranking, evaluation) using vector databases.
• Solid backend fundamentals: APIs, integrations, auth, and validation/debugging using tools like VS Code and Postman.
• Engineering discipline: modular design, testing, and CI/CD for scalable, production-grade delivery.

How we work

• Integrity: You tell the truth, you fix issues, you don’t ship for the sake of shipping.
• Curiosity: You keep learning, and you enjoy solving hard operational problems.
• Collaboration: You bring people with you, no heroics, no silos.
• Resilience: You stay calm when production gets tough.
• Customer-first: Every solution should make life easier and create a real soft landing.

Experience levels

Senior:

Minimum requirements

• Minimum: Matric and 6 - 10+ years engineering with ownership of architecture, reliability, and delivery outcomes
• Preferred: Degree/diploma + Azure AI Foundry exposure

Mid:

• Minimum: Matric and 3 - 5 years software engineering (Python-heavy), including production APIs/integrations
• Preferred: Degree/diploma + cloud exposure (Azure) + some CI/CD

Junior:

Minimum requirements

• Minimum: Matric + proof of coding competence (portfolio) in Python + basic API work
• Preferred: Diploma/Degree (CS/Eng/Maths/Stats/Data Science) or intensive software bootcamp + strong portfolio

Nice-to-haves

• MLOps / runtime: Docker, containerised deployments, automated lifecycle management.
• Familiarity with the Microsoft automation ecosystem (Power Automate, Copilot Studio) and CRM integrations (e.g., HubSpot/Salesforce).
• Exposure to call-centre integration patterns (e.g. ConnexAI) and real-time routing/agent assist environments.

Bottom line

We’re hiring an engineer who can turn models into reliable systems: deeply integrated, observable, secure, and genuinely useful. Not demos. Not science projects. Production.