Employer: Blacksmith Agency
Industry: Artificial Intelligence, AI Automation, Agentic Workflows & Software Engineering
Location: Remote (Emalahleni / Mpumalanga / Open to South Africa & Latin America)
Position Type: Full-Time Permanent
Department: Engineering
Role Focus: Production System Builder (Building & Shipping LLM Workflows, Not Academic Model Training)
Compensation: Market-Related Remote Software Engineering Salary
Posted Date: 24 October 2026
Closing Date: 24 November 2026
About Blacksmith Agency & The Machine Learning Engineer Role
Blacksmith Agency is a premier digital agency building an AI automation layer into every facet of its operational and delivery ecosystem.
Blacksmith Agency is recruiting a hands-on Machine Learning Engineer on a fully remote basis to own and ship its internal AI automation roadmap. This is a pragmatic, builder-focused role: you will not be training models from scratch, but rather building production systems, agentic pipelines, and custom integrations on top of leading Foundation Models (Anthropic Claude, OpenAI GPT-4) to reduce manual production hours across design, development, and CMS workflows.
Key Responsibilities & AI System Engineering Workflows
The Machine Learning Engineer will execute six core AI engineering deliverables:
1. Agentic Workflow Architecture
- Agent Systems: Build and deploy autonomous agentic workflows using frameworks like LangChain, LangGraph, CrewAI, or AutoGen.
- MCP Server Integration: Implement Model Context Protocol (MCP) server integrations as part of Blacksmith’s enterprise AI stack.
2. API Integration & Function Calling
- LLM APIs: Execute deep API integrations with Anthropic and OpenAI models, leveraging function calling, structured outputs, tool use, and real-time response streaming.
3. Operational Process Automation
- Pipeline Automation: Design and deploy automated data pipelines connecting WordPress, Figma, Google Workspace, and proprietary CMS platforms.
- Internal Tools: Build internal developer tools that drastically eliminate manual hours across web design and production pipelines.
4. Cross-Functional Delivery
- Roadmap Shipping: Collaborate directly with engineering and delivery leadership to transform documented workflow concepts into shipped, production-ready software.
Candidate Profile & Technical Requirements
To qualify for consideration, applicants must meet the following mandatory technical criteria:
- Software Engineering Foundation:
- Minimum 2+ years of experience building production software with Python or Node.js (Mandatory).
- LLM & Agent Framework Experience (CRITICAL):
- Proven hands-on experience building with LangChain, LangGraph, CrewAI, AutoGen, or similar multi-agent frameworks (Mandatory).
- Demonstrated portfolio or code examples showing shipped applications built on top of LLMs.
- API & Infrastructure Skills:
- Proven API integration work with OpenAI or Anthropic (function calling, tool use, streaming).
- Strong understanding of REST APIs, webhooks, JSON structures, and basic cloud deployment models.
- Communication & Remote Setup:
- Fluent English proficiency for asynchronous written communication and collaboration with US-based leadership.
- Ability to work independently from a secure home office in Emalahleni, Mpumalanga, or anywhere across South Africa.
Application Guidelines & Predictive Index Assessment
To ensure your application is processed, follow these exact steps:
- Submit Initial Link: Complete the short 2-minute online application form via the official Blacksmith Agency portal.
- Complete Assessment (CRITICAL): Your application will only be reviewed once you complete a short 6-minute Predictive Index (PI) assessment.
- Email Consistency: You must use the exact same email address for both the application form and the Predictive Index assessment.
Sample Selection & AI Engineering Interview Questions
Agentic Workflows & System Architecture
- Question (LangGraph State Management): “How do you handle cyclic graphs and state persistence in LangGraph when an agentic tool call fails or returns ambiguous JSON outputs from an LLM?”
- Question (Tool Use & Latency): “When connecting an LLM via function calling to external WordPress or Figma APIs, how do you optimize token usage and response latency for real-time production workflows?”
How to Apply
Apply online directly via the official Blacksmith Agency Careers link:
- Employer: Blacksmith Agency
- Location: Remote (Emalahleni / South Africa)
- Position: Machine Learning Engineer (AI Systems & Agentic Workflows)
- Closing Date: 24 November 2026