Remote AI Automation Engineer - Production AI Pipelines

Blacksmith Agency LLC; South Africa; Remote; ZAR 1,000,000 - 1,500,000; Full time

Benefits offered by this job

Remote work

Job summary

Blacksmith Agency is seeking a Machine Learning Engineer to own and ship the automation roadmap, building production systems on top of AI models. This builder role focuses on implementations rather than research, with hands-on work deploying integrations and internal tools that streamline design and development workflows.

You will work with LangChain, LangGraph and related frameworks, connect OpenAI/Anthropic APIs, and deliver automation pipelines across our stack while collaborating with a

Qualifications

• 2+ years building production software with Python or Node.js.
• Hands-on experience with LangChain, LangGraph, CrewAI, AutoGen.
• Proven API integration work with OpenAI or Anthropic (function calling, tool use, streaming).
• Strong understanding of REST APIs, webhooks and basic cloud deployment.
• English sufficient for async written work with a US-based team.

Responsibilities

• Own and ship internal automation roadmap and production systems.
• Build production software on top of AI models and services.
• Implement agent workflows using LangChain/LangGraph and related tools.
• Create integrations with OpenAI and Anthropic APIs (tooling and calls).
• Develop internal tools to reduce manual hours across processes.

Skills

Python

Node.js

LangChain

LangGraph

CrewAI

AutoGen

OpenAI API

Anthropic API

REST APIs

Cloud deployment

English communication

Job description

Blacksmith Agency is seeking a Machine Learning Engineer to own and ship the automation roadmap, building production systems on top of AI models. This builder role focuses on implementations rather than research, with hands-on work deploying integrations and internal tools that streamline design and development workflows.

You will work with LangChain, LangGraph and related frameworks, connect OpenAI/Anthropic APIs, and deliver automation pipelines across our stack while collaborating with a