H
Lead AI Engineer at Hire Resolve
Hire Resolve
September 12, 2026
Full-time
On-site
Job Description
A leading South African packaging manufacturer that specializes in producing corrugated packaging and Single Face Kraft (SFK) products is seeking a Lead AI Engineer who will be the technical owner of an internally built agentic system that currently runs the company's planning cycle end-to-end.
Responsibilities:
Production Governance: Own existing agentic MRP/planning systems; establish source control, managed environments, clear documentation, and formal deployment pipelines.
Risk & Fallback Management: Eliminate single points of failure (technical and operational). Maintain tested rollbacks, kill-switches, and documented manual fallback plans.
Bounded Autonomy & Safety: Program hard limits on agent scope/financial values, enforce human-in-the-loop approval gates for irreversible actions, and ensure loud failures over silent errors.
Eval Harnesses: Build and run historical evaluation test suites against all prompt, code, or model updates prior to deployment.
Staged Releases: Enforce a strict rollout pathway (Shadow Running $\rightarrow$ Pilot $\rightarrow$ Canary $\rightarrow$ Earned Autonomy).
Observability & Auditing: Implement complete telemetry (inputs, reasoning logs, actions) and queryable audit trails that satisfy financial and governance standards.
Process Discovery: Engage directly with factory floor and business operations to understand actual workflows before designing automation.
End-to-End Delivery: Scope, architect, and build scalable systems across planning, logistics, procurement, and finance—recommending "go/no-go" based on pilot metrics.
Commercial Controls: Ensure financial actions adhere to CFO/external audit requirements (segregation of duties, approval limits).
Tooling & AI-Assisted Dev: Leverage modern AI dev tools (e.g., Cursor, Claude Code) as primary workflows with rigorous code review and testing standards.
Vendor & Cost Management: Direct model selection, transaction cost optimization, data residency, and vendor lock-in prevention.
Enablement & Reporting: Provide runbooks and training for business users; report directly to executive leadership on progress, risk, and technical debt.
Minimum Requirements:
Production Release Discipline: Proven track record of owning commercial software running in live production (including staging, canary rollouts, incident handling, and rollbacks). Prototypes or internal tools only will not be shortlisted.
End-to-End System Shipping: Demonstrable history of delivering full-lifecycle software solutions, including error handling, monitoring, and handover.
Engineering Background: Minimum 8+ years in software engineering, with 3+ years in a Senior, Lead, or Principal role. Strong software engineering fundamentals take precedence over pure ML research.
AI-Assisted Workflows: High fluency with AI-assisted development tools (Cursor, Claude Code, or equivalent), assessed via live practical evaluation.
Desirable Experience
Practical experience building with LLMs and agentic workflows (prompt engineering, tool use, evaluation frameworks, handling non-determinism).
Familiarity with ERP, MRP, or manufacturing supply chain environments.
Willingness to master corrugated packaging and manufacturing process nuances.
A leading South African packaging manufacturer that specializes in producing corrugated packaging and Single Face Kraft (SFK) products is seeking a Lead AI Engineer who will be the technical owner of an internally built agentic system that currently runs the company's planning cycle end-to-end.
Responsibilities:
Production Governance: Own existing agentic MRP/planning systems; establish source control, managed environments, clear documentation, and formal deployment pipelines.
Risk & Fallback Management: Eliminate single points of failure (technical and operational). Maintain tested rollbacks, kill-switches, and documented manual fallback plans.
Bounded Autonomy & Safety: Program hard limits on agent scope/financial values, enforce human-in-the-loop approval gates for irreversible actions, and ensure loud failures over silent errors.
Eval Harnesses: Build and run historical evaluation test suites against all prompt, code, or model updates prior to deployment.
Staged Releases: Enforce a strict rollout pathway (Shadow Running $\rightarrow$ Pilot $\rightarrow$ Canary $\rightarrow$ Earned Autonomy).
Observability & Auditing: Implement complete telemetry (inputs, reasoning logs, actions) and queryable audit trails that satisfy financial and governance standards.
Process Discovery: Engage directly with factory floor and business operations to understand actual workflows before designing automation.
End-to-End Delivery: Scope, architect, and build scalable systems across planning, logistics, procurement, and finance—recommending "go/no-go" based on pilot metrics.
Commercial Controls: Ensure financial actions adhere to CFO/external audit requirements (segregation of duties, approval limits).
Tooling & AI-Assisted Dev: Leverage modern AI dev tools (e.g., Cursor, Claude Code) as primary workflows with rigorous code review and testing standards.
Vendor & Cost Management: Direct model selection, transaction cost optimization, data residency, and vendor lock-in prevention.
Enablement & Reporting: Provide runbooks and training for business users; report directly to executive leadership on progress, risk, and technical debt.
Minimum Requirements:
Production Release Discipline: Proven track record of owning commercial software running in live production (including staging, canary rollouts, incident handling, and rollbacks). Prototypes or internal tools only will not be shortlisted.
End-to-End System Shipping: Demonstrable history of delivering full-lifecycle software solutions, including error handling, monitoring, and handover.
Engineering Background: Minimum 8+ years in software engineering, with 3+ years in a Senior, Lead, or Principal role. Strong software engineering fundamentals take precedence over pure ML research.
AI-Assisted Workflows: High fluency with AI-assisted development tools (Cursor, Claude Code, or equivalent), assessed via live practical evaluation.
Desirable Experience
Practical experience building with LLMs and agentic workflows (prompt engineering, tool use, evaluation frameworks, handling non-determinism).
Familiarity with ERP, MRP, or manufacturing supply chain environments.
Willingness to master corrugated packaging and manufacturing process nuances.