AI AGENTS ENGINEER

Role Summary

Mindtrace is hiring an Agentic Workflow / AI Product Engineer to build the agentic workflow layer for its inspection products. This person will help create agents that make industrial AI systems easier to deploy, diagnose, explain, support, and improve.

In this role, you will build production-quality agentic product workflows around inspection evidence, deployment state, hardware health, logs, model outputs, customer workflows, support history, and field feedback.

The ideal candidate has prior hands-on experience building agents or agentic workflows, preferably in real production software rather than only prototypes. They should be fluent in Python and comfortable building robust AI applications that use tools, retrieve context, reason over multimodal evidence, and operate within clear human-in-the-loop boundaries.

What You Will Do

• Build agentic workflows for Mindtrace inspection products, including deployment support, system health, incident triage, inspection explanation, reporting, and data-curation loops.
• Develop agents that can use tools, inspect logs, retrieve relevant context, summarize evidence, recommend next actions, and package issues for escalation.
• Work on agentic product workflows that support deployment, inspection review, diagnostics, reporting, escalation, and continuous improvement.
• Integrate agents with product APIs, databases, image records, model outputs, configuration state, hardware status, support history, and customer-specific metadata.
• Design agent workflows that sit inside real task surfaces rather than generic chatbot experiences.
• Build evaluation and observability for agent behaviour, including tool-call history, task success metrics, regression tests, failure modes, traceability, and guardrails.
• Collaborate with ML, product, hardware, client, and deployment teams to turn field experience into reusable product knowledge.
• Help define where agents should explain, recommend, summarize, prepare actions, or escalate, while preserving clear human authority for consequential production decisions.
• Prototype quickly, but harden the right workflows into maintainable product features.

Required Skills and Experience

• Prior hands-on experience building agents, agentic workflows, LLM applications, or AI workflow products.
• Strong Python fluency for application development, workflow orchestration, AI integration, testing, and data processing.
• Experience with LLM or VLM application patterns such as tool use, structured outputs, retrieval, memory/context management, prompt design, and evaluation.
• Strong backend or full-stack product engineering skills, including APIs, databases, events, permissions, workflow state, and integration with existing systems.
• Ability to build reliable human-in-the-loop AI workflows where agent actions are observable, reviewable, and bounded.
• Experience designing or implementing evaluation for AI systems, including regression testing, task success metrics, failure analysis, and reliability checks.
• Comfort working with multimodal operational data, including images, logs, model outputs, metadata, configuration state, and user feedback.
• Clear communication and product judgement: able to turn ambiguous operational needs into useful workflows that real users can trust.

Highly Beneficial Skills

• Experience shipping agents or LLM applications in production environments.
• Experience with vision-language models or multimodal AI over images, video, inspection evidence, diagrams, screenshots, or technical records.
• Experience with RAG over structured and unstructured data.
• Familiarity with observability, log analysis, incident response, support tooling, runbooks, or developer tools.
• Experience building workflow UIs where AI assists a specific operational task.
• Familiarity with manufacturing, quality inspection, industrial automation, field support, or regulated operational environments.
• Experience with vector databases, embeddings, search systems, or knowledge retrieval.
• Experience with model governance, audit trails, release control, permissions, or safety-sensitive AI workflows.
• Familiarity with OpenAI, Anthropic, Gemini, open-source LLM/VLM stacks, LangGraph-style orchestration, or similar agent frameworks. The candidate should understand enough vision and inspection context to collaborate well with ML and hardware specialists, while remaining primarily focused on agentic workflows and AI product engineering.

Candidate Profile

The strongest candidates will have direct experience building and deploying agentic systems into industrial, operational, or similarly complex production environments. They should know how to move beyond prototypes into reliable agent workflows that save time, reduce deployment friction, explain evidence, improve support workflows, and make inspection systems easier to operate and trust. This person should be able to answer questions like:

• What should this agent be allowed to do, and where should a human remain in control?
• What context does the agent need to diagnose this issue reliably?
• How do we know the agent is working, and how do we detect when it fails?
• How do we turn one deployment issue into reusable product knowledge?
• How should agent activity be logged, inspected, evaluated, and improved over time?

We are an equal opportunity employer. We are committed to equal employment opportunity regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, citizenship, marital status, disability, gender identity or Veteran status.