Ricardo Elias Roche
Community@ricardoroche · Brooklyn, NY
ML/AI Engineer
Agent Skills by Ricardo Elias Roche
Showing 29 vetted skills indexed across 1 GitHub repositories.
pytest-patterns
Enforce pytest patterns for fixtures, parametrize, mocking, and async tests.
docstring-format
Enforce Google-style docstrings for Python functions and classes.
evaluation-metrics
Apply standardized metrics to LLM evaluation datasets for reproducible assessments.
llm-app-architecture
Standardize asynchronous LLM API calls with streaming, retries, and token management.
observability-logging
Automate JSON-structured logging and OpenTelemetry tracing for AI/LLM pipelines.
git-workflow-standards
Enforce conventional commits, branch naming, PR templates, and semantic versioning.
performance-profiling
Profile Python CPU, memory, and async performance with cProfile and memory_profiler.
monitoring-alerting
Automate Prometheus/Grafana monitoring, alerting, and SLO/SLI definitions.
ai-security
Detect prompt injection attempts in AI and LLM applications.
tool-design-pattern
Enforce standardized AI tool design patterns with Pydantic schemas and error handling.
pii-redaction
Redact PII from logs, error messages, and debug output.
agent-orchestration-patterns
Standardize Pydantic tool schemas and agent state for multi-agent orchestration.
openspec-authoring
Enforce OpenSpec authoring conventions for proposals, tasks, and spec files.
pydantic-models
Enforce Pydantic model patterns for API data validation and serialization.
docs-style
Enforce voice, clarity, and navigation patterns in documentation drafts.
spec-templates
Generate Markdown templates for proposals, design docs, ADRs, READMEs, and changelogs.
code-review-framework
Apply a structured review framework to Python pull requests.
python-packaging
Automate Python packaging configuration in pyproject.toml for PyPI distribution.
rag-design-patterns
Standardizes chunking, vector database integration, embedding management, and retrieval reranking for RAG pipelines.
type-safety
Enforce comprehensive type hints in Python code with mypy-compatible signatures.
model-selection
Automate LLM model and provider selection with registry-driven configuration.
fastapi-patterns
Enforces FastAPI endpoint patterns including dependency injection, authentication, validation, and error handling for API definitions.
dependency-management
Automate Python dependency management with uv, Poetry, or pip.
prompting-patterns
Automate prompt engineering with structured templates, few-shot examples, and injection prevention.