Black Belt 1st Dan · Dan exam · AGE
Agent Engineer
Validates that you can design long-running, multi-step agents and operate them with tools, memory, evaluation and observability in place, keeping failures and costs under control.
Daily practice · Free
10 questions · 25 min · No sign-up needed
Dan exam · Proctored
40 questions · 70 min · Raw score of 75 or more · $3.99 / one attempt
Dan exam detailsExam overview
- Questions
- 40
- Time
- 70 min
- Pass mark
- 75 pts
- Recommended experience
- Red Belt (LLM API Developer) level and at least 6 months of experience operating agents
Every question has five choices and is marked on an OMR answer sheet. Multiple-choice and true-or-false questions ask for the right combination from a box; ordering and matching questions ask for the right arrangement. Judgment questions (L3) are worth 3 points and all others 2; the raw score is converted to a 100-point scale and graded 1 to 9.
Exam domains
1. Agent architecture20%
2. Tools and MCP20%
3. Context, memory and state20%
4. Evaluation and observability20%
5. Operations, cost and safety20%
Weights follow the Dan exam. Practice and mock exams may mix domains differently.
What this exam verifies
- Choose the workflow or agent structure that fits the task
- Design tools and MCP servers within safe permission boundaries
- Manage context, memory and state for long-running work
- Catch quality regressions with evaluation and observability
- Turn failure handling, cost control and human approval points into operating policy
Who it is for
- Engineers who design and operate agent systems
- AI platform team leads
Does not cover model pre-training.
Domain details
Domain 1. Agent architecture
- Knowledge: Workflows vs. agents, Routing, parallelization and orchestrator-workers, The cost of multi-agent systems
- Skills: Start with the simplest structure
Domain 2. Tools and MCP
- Knowledge: Tool design principles, MCP servers and clients, Permissions and approvals, Managing tool result size
- Skills: Define tools that agents find hard to misuse, Grant least privilege
Domain 3. Context, memory and state
- Knowledge: Context compaction, External memory, Checkpoints, Handing off long-running work
- Skills: Prevent context pollution in long runs, Design for resuming after an interruption
Domain 4. Evaluation and observability
- Knowledge: Trajectory evaluation, Outcome-based evaluation, Tracing, Regression sets
- Skills: Analyze agent failures step by step, Make deployment decisions with evaluation metrics
Domain 5. Operations, cost and safety
- Knowledge: Budget limits, Timeouts and retries, Human approval points, Sandboxes
- Skills: Stop runaway runs and infinite loops, Require approval for irreversible actions
Official documentation
Black Belt Course · Agent Engineer
Practice, mock exams and every course book are free. Pay only when you take an official test.