Intermediate AI Engineer

Salt; South Africa; In loco; ZAR 900.000 - 1.400.000; Tempo pieno

Descrizione del lavoro

Salt is seeking AI Engineers to drive from problem understanding to deployment, evaluation, and real-world use. You’ll work with LLMs and agentic AI, integrating tools, APIs and enterprise data to create robust, production-ready systems.

Expect to tackle multi-step workflows, memory and tool-calling patterns, and advanced tooling like LangChain, LangGraph, and Copilot as part of the build. You’ll be measured by the impact of your AI systems when used by real users, not just demos.

Competenze

• Experience designing and deploying AI systems into real-world use cases.

Mansioni

• Take AI from problem framing to production deployment and ongoing improvement.

Conoscenze

RAG

Retrieval

Vector search

Agents

Multi-step workflows

Context memory

Structured tool calling

LangChain

LangGraph

Google ADK

AutoGen

MCP

Guardrails

Software engineering

Production deployment

Claude Code

Cursor

Copilot

Strumenti

LangChain

LangGraph

Google ADK

AutoGen

Claude Code

Cursor

Copilot

Descrizione del lavoro

AI Engineers Have you built an AI agent that actually made it past the demo? We are headhunting AI Engineers: You'll be taking AI all the way into the real world, from understanding the problem and experimenting with an approach, through to building, deploying, evaluating and improving systems that people actually use.

Think: LLMs and agentic AI

• RAG, retrieval and vector search


• Agents that can use tools, APIs and enterprise data


• Multi-step / multi-agent workflows


• Context, memory and structured tool calling


• LangChain, LangGraph, Google ADK, AutoGen or similar


• MCP and modern agent-tool integrations


• Testing, evaluation, tracing and guardrails


• Strong software engineering and production deployment


• Using tools like Claude Code, Cursor or Copilot as part of how you build

You don't need to have every framework above. What matters more is that you can point to something and say: \"I built this. I understand how it works. I put it into production. And I know what happened when real users started using it.\" If you're building with LLMs, agents and modern AI engineering and want your next role to go deeper than another chatbot or internal POC,