Agentic Engineer (Entry level) at Deloitte

Job Description

An Agentic Engineer designs, builds, and manages systems where AI agents can perceive, reason, act, and learn autonomously to achieve goals. An Agentic Engineer:

• Develops architectures that define how agents gather context, make decisions, coordinate with other agents, and interact with external tools or APIs.
• Designs and build robust runtime environments for agents, which form the foundation for the cognition process.
• Designs and implements appropriate security, observability, and governance measures to supplement the runtime environment.
• Implements reasoning and control frameworks (e.g., planning, reflection loops, evaluation pipelines) that allow agents to operate safely and intelligently.
• Optimises human–AI workflows, ensuring agents augment human capability rather than create friction or risk.
• Monitors, debugs, and improve agent behaviour, using telemetry and evaluation signals to refine performance over time.
• Ensures safety, reliability, and alignment, establishing guardrails and governance for autonomous operations.

Qualifications

Role profile

An Agentic Engineer designs, builds, and manages systems where AI agents can perceive, reason, act, and learn autonomously to achieve goals. An Agentic Engineer:

• Develops architectures that define how agents gather context, make decisions, coordinate with other agents, and interact with external tools or APIs.
• Designs and build robust runtime environments for agents, which form the foundation for the cognition process.
• Designs and implements appropriate security, observability, and governance measures to supplement the runtime environment.
• Implements reasoning and control frameworks (e.g., planning, reflection loops, evaluation pipelines) that allow agents to operate safely and intelligently.
• Optimises human–AI workflows, ensuring agents augment human capability rather than create friction or risk.
• Monitors, debugs, and improve agent behaviour, using telemetry and evaluation signals to refine performance over time.
• Ensures safety, reliability, and alignment, establishing guardrails and governance for autonomous operations.

Required Qualifications and Experience

This is an entry-level position and work experience is not a pre-requisite. The following requirements do however apply:

• A bachelor’s degree. Based on the information presented in “Agentic Engineering in the Deloitte context”, a wide variety of degrees will be considered.
• A demonstrable interest in Generative AI and Agentic AI. This interest can be demonstrated through a wide variety of means, examples of which include courses completed, certifications obtained, projects completed, articles written, books and research papers read etc.
• Building AI systems requires programming and proficiency in Python is required.
• Knowledge- and experience of relevant protocols (e.g. MCP, ACP, A2A, ANP etc.) and platforms/frameworks (e.g. LangChain, LangGraph, LlamaIndex, CrewAI, AutoGen etc.) is beneficial.
• Previous experience with- and certifications in AWS, Azure and GCP is also beneficial. AI and Generative AI certifications are particularly valuable.

Important behavioural traits

General:

• Excellent English communication skills, both written and verbal.
• Effective interpersonal and relationship-building skills.
• Adaptable, managing change and ambiguity with ease.
• Effective problem-solving ability.

Role specific:

• Systems thinking. You view problems as interconnected systems rather than isolated components.
• Bias for experimentation and iteration. You like to prototype, observe, gather data and refine.
• Reflective and meta-cognitive thinking. You enjoy practicing second-order and meta-level reasoning.
• Clarity in instruction and decomposition. You are exceptional at breaking down tasks, specifying steps, defining boundaries and writing instructions that are unambiguous but flexible.
• Curiosity and cross-disciplinary learning. You enjoy learning in- and connecting many domains e.g., AI and ML, distributed systems, human factors, psychology, workflow design, product thinking etc.
• Ambiguity and emergent behaviour. You are comfortable navigating uncertainty and shaping it into predictable performance.
• Builder mentality. You enjoy building, testing, and improving working systems.