AI Agents
AI Agents 1
Fifteen implementable exercises drawn from the concrete algorithms behind agent engineering: context trimming, reciprocal rank fusion, memory eviction, tool-schema validation, parallel-call scheduling, circuit breakers, ReWOO plan resolution, tree-of-thoughts beam search, token cost accounting and pass@k scoring.
50
problems
5
topics
15
free
01
Agent Loop Control & Planning
10 problems- 1Audit an Agent Trace for Step Limits and LoopsFREEEasy
- 2Detect a Stuck Agent LoopPROEasy
- 3Should the Agent Loop ContinuePROEasy
- 4Resolve a ReWOO Plan into Execution WavesFREEMedium
- 5Resolve a Plan DAG into Parallel WavesPROMedium
- 6Reflexion Retry DecisionPROMedium
- 7Greedy Subgoal Scheduling Under a Step BudgetPROMedium
- 8Detect Repeated State Cycles in a TrajectoryPROMedium
- 9Beam Search over a Tree of ThoughtsFREEHard
- 10Beam Search over Partial PlansPROHard
02
Context & Agent Memory
10 problems- 11Trim a Message History to a Token BudgetFREEEasy
- 12Estimate Token Count of a Message ListPROEasy
- 13Sliding Window Keeping the System MessagePROEasy
- 14Reciprocal Rank Fusion over Hybrid RetrieversFREEMedium
- 15Deduplicate and Evict Agent MemoriesPROMedium
- 16Trigger Summarization at a Budget FractionPROMedium
- 17Recency and Relevance Memory ScoringPROMedium
- 18Chunk Text into Overlapping WindowsPROMedium
- 19Maximal Marginal Relevance SelectionPROMedium
- 20Priority Context Packing Under a Token BudgetFREEHard
03
Production: Cost, Caching & Evals
10 problemsSign up free to see all 50 problems
Create a free account to see the full AI Agents 1 list, track your progress, and save problems to lists.
Sign up free