Traigent Learning Roadmap¶
A concise path from first run to production-ready evaluations.
Tracks¶
- Beginner (60-90 min): mock-mode runs, core knobs.
- Builder (2-3 hours): balance cost/latency/accuracy, add safety, ship to CI.
- Researcher (3-4 hours): advanced playbooks, metrics, RAG evaluation.
Week 1: Foundations (beginner)¶
- Sanity check:
python examples/core/simple-prompt/run.py - RAG toggle:
TRAIGENT_MOCK_LLM=true python examples/core/rag-optimization/run.py - Few-shot prompts:
TRAIGENT_MOCK_LLM=true python examples/core/few-shot-classification/run.py - Trade-offs:
TRAIGENT_MOCK_LLM=true python examples/core/multi-objective-tradeoff/run_anthropic.py
Outcome: you can run examples, read metrics, and tweak configuration spaces.
Week 2: Applied patterns (builder)¶
- Control spend/latency:
examples/core/token-budget-summarization/run.py - Quality and tone:
examples/core/prompt-style-optimization/run.py - Safety:
examples/core/safety-guardrails/run.py->examples/advanced/ai-engineering-tasks/p1_safety_guardrails/ - Structured outputs:
examples/core/structured-output-json/run.py - Tool calling:
examples/core/tool-use-calculator/run.py
Outcome: you can constrain costs, add guardrails, and tune prompts.
Week 3: Advanced exploration (researcher)¶
- Optimization routing:
examples/advanced/optimization-routing/(local search and no-egress examples) - Context engineering & RAG eval:
examples/advanced/ai-engineering-tasks/p0_context_engineering/andexamples/advanced/ragas/ - Metrics and analysis:
examples/advanced/results-analysis/andexamples/advanced/metric-registry/ - CI/CD integration:
examples/integrations/ci-cd/
Outcome: you can experiment with new metrics, evaluate RAG systems, and integrate Traigent into pipelines.
Running and debugging¶
pip install "traigent[recommended]"
# In local tutorial code, call enable_mock_mode_for_quickstart() before running LLM examples.
ls examples/datasets/<example-name>
rg "optimize(" examples
TRAIGENT_LOG_LEVEL=DEBUG.