Main Page
27,667
Total Pages
1,118
Workflows
7,697
Principles
16,832
Implementations
1,245
Heuristics
909
Environments
Welcome to Leeroopedia
Your ML & Data Knowledge Wiki. Best practices and expert-level knowledge for Machine Learning and Data Engineering, covering 1000+ frameworks and libraries from training to deployment.
Browse implementation patterns, configuration guides, debugging heuristics, and battle-tested defaults for frameworks like vLLM, DeepSpeed, Megatron-LM, FlashAttention, Triton, Unsloth, LangChain, and many more. Every page is structured so both humans and AI agents can find what they need fast.
Connect your AI coding agent. Plug Leeroopedia into your favorite coding agent, and let it build robust AI/ML systems autonomously:
- SuperML plugin — converts your AI coding agent into an expert ML engineer with agentic memory
- Leeroopedia MCP — search over best-practices and skills of ML/AI
- Kapso — experimentation platform for autonomous AI/ML software building
Browse by Category
| Category | Description | Browse |
|---|---|---|
| Workflows | Step-by-step processes and procedures | Browse All |
| Principles | Core ideas and foundational knowledge | Browse All |
| Implementations | Code-level details and modules | Browse All |
| Heuristics | Best practices and guidelines | Browse All |
| Environments | Setup and configuration guides | Browse All |
Explore Pages
Workflows
- Workflow:Openai Evals Building a custom eval
- Workflow:Apache Paimon Blob Storage With Descriptors
- Workflow:Duckdb Duckdb Source Amalgamation And Packaging
- Workflow:Apache Spark Release Process
- Workflow:Fede1024 Rust rdkafka Mock Cluster Testing
- Workflow:NVIDIA NeMo Aligner RLHF PPO Training
- Workflow:ClickHouse ClickHouse Server Deployment
- Workflow:Microsoft Semantic kernel Kernel Setup And Chat Completion
- Workflow:Apache Shardingsphere Metadata DDL Refresh
- Workflow:Haotian liu LLaVA LoRA Finetuning
Principles
- Principle:ArroyoSystems Arroyo Worker Scheduling
- Principle:Datahub project Datahub Emitter Lifecycle
- Principle:Haifengl Smile File Data Loading
- Principle:InternLM Lmdeploy W8A8 Quantized Inference
- Principle:Togethercomputer Together python Endpoint Management
- Principle:NVIDIA DALI Generic Iterator
- Principle:Langgenius Dify Docker Infrastructure
- Principle:Huggingface Diffusers Prior Preservation Training
- Principle:Apache Dolphinscheduler Frontend Dependency Management
- Principle:CrewAIInc CrewAI Managed Execution
Implementations
- Implementation:OpenGVLab InternVL Classification Main
- Implementation:Microsoft Onnxruntime CUDA TrainingSplit
- Implementation:Huggingface Open r1 SFTTrainer Usage
- Implementation:CARLA simulator Carla Draw Skeleton Example
- Implementation:Infiniflow Ragflow AvatarUpload Component
- Implementation:Mage ai Mage ai Datadog Streams
- Implementation:Recommenders team Recommenders Notebook Utils
- Implementation:Marker Inc Korea AutoRAG Runner Run
- Implementation:Helicone Helicone MapperBuilder PathMapper
- Implementation:Spotify Luigi ExternalDailySnapshot
Heuristics
- Heuristic:Dagster io Dagster Lazy Import Pattern
- Heuristic:Deepspeedai DeepSpeed Sequence Parallel PyTorch Version
- Heuristic:Haotian liu LLaVA Tokenizer Version Offset Correction
- Heuristic:PeterL1n BackgroundMattingV2 Refine Mode Selection
- Heuristic:Huggingface Optimum GPTQ Quantization Defaults
- Heuristic:Open compass VLMEvalKit Judge Model Selection By Dataset
- Heuristic:Diagram of thought Diagram of thought Typed Records For Auditability
- Heuristic:Mit han lab Llm awq GPU Memory Management Patterns
- Heuristic:LaurentMazare Tch rs MPS Weight Loading Workaround
- Heuristic:MaterializeInc Materialize CI Retry Strategies
Environments
- Environment:Huggingface Datasets TensorFlow Integration
- Environment:Sgl project Sglang Distributed
- Environment:Vllm project Vllm Environment Variables
- Environment:Ggml org Llama cpp Python Conversion Environment
- Environment:Nightwatchjs Nightwatch Selenium WebDriver 4
- Environment:Cohere ai Cohere python Python SDK Runtime
- Environment:Microsoft Onnxruntime CUDA GPU Environment
- Environment:Apache Flink Java Build Environment
- Environment:NVIDIA TransformerEngine GPU Compute Capability
- Environment:Google research Deduplicate text datasets Python TFDS Environment