home / skills / orchestra-research / ai-research-skills

orchestra-research/ai-research-skills

Comprehensive open-source library of AI research and engineering skills for any AI model. Package the skills and your claude code/codex/gemini agent will be an AI research agent with full horsepower. Maintained by Orchestra Research.

86 skills

Sponsored

Advertise on playbooks

Put your product in front of developers browsing MCP servers and agent skills. ↗

orchestra-research/ai-research-skills

This skill optimizes LLM inference on NVIDIA GPUs with TensorRT for maximum throughput and lowest latency in production.

orchestra-research/ai-research-skills

This skill helps you run interpretability experiments on massive PyTorch models remotely using nnsight and optional NDIF, with a single code path.

orchestra-research/ai-research-skills

This skill helps you apply expert content moderation with LlamaGuard to filter unsafe prompts and outputs across chat systems.

orchestra-research/ai-research-skills

This skill detects prompt injections and jailbreak attempts in LLM apps, ensuring safer interactions and reliable third-party data filtering.

orchestra-research/ai-research-skills

This skill provides expert guidance for distributed training with DeepSpeed, covering ZeRO, pipeline parallelism, FP16/BF16/FP8, and optimization best

Advertise on playbooks

Reach thousands of developers actively evaluating dev tools, MCP servers, and agent skills. Floating banners, pinned cards, and in-feed placements.

orchestra-research/ai-research-skills

This skill enables zero-shot image understanding and cross-modal retrieval using CLIP for image search, moderation, and vision-language tasks.

orchestra-research/ai-research-skills

This skill helps you deploy large language models with 4-bit AWQ quantization, delivering 2.5-3x faster inference while preserving accuracy.

lm-evaluation-harness

orchestra-research/ai-research-skills

This skill benchmarks LLMs across 60+ tasks, helping you compare models, track progress, and report standardized academic results efficiently.

orchestra-research/ai-research-skills

This skill helps you manage end-to-end ML lifecycles with MLflow, including experiment tracking, model registry, and reproducible deployments across frameworks.

orchestra-research/ai-research-skills

This skill helps you visualize training metrics, debug models, compare experiments, and profile performance with TensorBoard.

orchestra-research/ai-research-skills

This skill helps you rapidly build LLM applications using agents, chains, and RAG with multi-provider support and robust tooling.

orchestra-research/ai-research-skills

This skill enables fast billion-scale vector similarity with FAISS, guiding deployment, index selection, and GPU-accelerated search for high-performance

sentence-transformers

orchestra-research/ai-research-skills

This skill helps generate high-quality embeddings for semantic search and retrieval using sentence-transformers, enabling efficient RAG, clustering, and

orchestra-research/ai-research-skills

This skill helps you build complex AI systems with declarative LM programming, automatic prompt optimization and modular RAG pipelines for reliable outputs.

20-ml-paper-writing

orchestra-research/ai-research-skills

This skill helps you draft publication-ready ML papers for top conferences by providing proactive drafting, citation verification, LaTeX templates, and

stable-diffusion

orchestra-research/ai-research-skills

This skill helps you generate high-quality images from text prompts, perform image-to-image tasks, and optimize diffusion workflows with Stable Diffusion.

orchestra-research/ai-research-skills

This skill helps you learn transformer basics by guiding you through nanoGPT style GPT-2 reproduction, training, and experimentation for educational purposes.

orchestra-research/ai-research-skills

This skill helps you implement and train LLMs with LitGPT across 20+ pretrained architectures for clean, production-ready workflows.

orchestra-research/ai-research-skills

This skill helps you deploy and experiment with Mamba selective state-space models for efficient linear-time sequence processing on GPUs.

orchestra-research/ai-research-skills

This skill helps you optimize sequence processing with RWKV by combining parallel training and sequential inference for long-context models.

orchestra-research/ai-research-skills

This skill enables scalable pretraining of large language models using PyTorch Torchtitan 4D parallelism across GPUs, delivering faster training with efficient

huggingface-tokenizers

orchestra-research/ai-research-skills

This skill accelerates tokenization and tokenizer training with HuggingFace Tokenizers, delivering ultra-fast performance and alignment support for NLP

orchestra-research/ai-research-skills

This skill provides expert guidance for fine-tuning LLMs with Axolotl, including YAML configs, 100+ models, and multimodal support.

orchestra-research/ai-research-skills

Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.

orchestra-research/ai-research-skills

This skill provides expert guidance for fast fine-tuning with Unsloth, enabling 2-5x training speed and reduced memory usage.

orchestra-research/ai-research-skills

This skill helps you train and analyze Sparse Autoencoders with SAELens to extract interpretable, monosemantic features from neural activations.

grpo-rl-training

orchestra-research/ai-research-skills

This skill guides GRPO/RL fine-tuning with TRL to enforce formats, optimize reasoning, and align domain-specific behaviors using reward-based training.

orchestra-research/ai-research-skills

This skill guides enterprise RL training with miles for large MoE models, enabling FP8/INT4, train-inference alignment, and speculative RL for throughput.

orchestra-research/ai-research-skills

This skill helps you accelerate RL-based LLM post-training with slime's Megatron-LM and SGLang for scalable data generation and rollout.

orchestra-research/ai-research-skills

This skill helps you accelerate RL experimentation with PyTorch-native abstractions, enabling clean algorithm isolation and scalable training using torchforge.

orchestra-research/ai-research-skills

This skill guides reinforcement learning based training of large language models using verl across PPO, GRPO, and other RL algorithms.

constitutional-ai

orchestra-research/ai-research-skills

This skill helps you align AI safety using self-critique and AI feedback, reducing harmful outputs without human labeling.

nemo-guardrails

orchestra-research/ai-research-skills

This skill enforces runtime safety for LLMs with configurable jailbreaking, toxicity, PII, and fact-checking rails to improve reliability.

orchestra-research/ai-research-skills

This skill helps you optimize large-scale LLM training with Megatron-Core, enabling efficient 2B-462B parameter models using advanced parallelism.

orchestra-research/ai-research-skills

This skill helps you integrate PyTorch FSDP2 into training scripts with correct initialization, sharding, mixed precision, and DTensor-based checkpointing.

pytorch-lightning

orchestra-research/ai-research-skills

This skill helps you streamline PyTorch Lightning training, automate distributed execution, and reduce boilerplate for scalable, reproducible experiments.

orchestra-research/ai-research-skills

This skill helps you manage Lambda Labs GPU Cloud resources for scalable ML training and inference with persistent storage and easy SSH access.

orchestra-research/ai-research-skills

This skill guides you to run ML workloads on Modal's serverless GPU platform with automatic scaling and on-demand pricing.

orchestra-research/ai-research-skills

This skill helps orchestrate ML workloads across multiple clouds with automatic cost optimization and spot instance recovery.

orchestra-research/ai-research-skills

This skill helps you deploy AI models efficiently on consumer hardware using GGUF quantization for flexible 2-8 bit inference.

flash-attention

orchestra-research/ai-research-skills

This skill accelerates transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction on long sequences.

orchestra-research/ai-research-skills

This skill helps you compress large language models to 4-bit precision with minimal accuracy loss, enabling faster inference and smaller memory footprints.

orchestra-research/ai-research-skills

This skill enables calibration-free 4/3/2-bit model quantization using HQQ for fast, deployment-ready quantization with HuggingFace and vLLM.

orchestra-research/ai-research-skills

This skill enables efficient LLM inference on CPU and non-NVIDIA hardware, enabling edge deployment and Apple Silicon performance with GGUF quantization.

orchestra-research/ai-research-skills

This skill helps deploy high-throughput LLM serving with vLLM, enabling OpenAI-compatible endpoints, quantization, and tensor parallelism for production

orchestra-research/ai-research-skills

This skill helps you benchmark LLMs across 100+ benchmarks with containerized, scalable evaluation on local Docker, Slurm HPC, or cloud platforms.

orchestra-research/ai-research-skills

This skill helps you enforce structured generation with regex and grammars, guaranteeing valid JSON/XML/code and guiding multi-step workflows.

orchestra-research/ai-research-skills

This skill merges multiple fine-tuned models with mergekit to create specialized, higher-performing models without retraining.

orchestra-research/ai-research-skills

This skill helps you compress large language models with pruning techniques like Wanda and SparseGPT, achieving 40-60% size reduction with minimal accuracy

orchestra-research/ai-research-skills

This skill helps you implement language-independent tokenization with SentencePiece to support multilingual models and reproducible vocabularies.

speculative-decoding

orchestra-research/ai-research-skills

This skill accelerates LLM inference using speculative decoding, Medusa heads, and lookahead techniques to boost speed and reduce latency.

orchestra-research/ai-research-skills

This skill helps you optimize alignment training with SimPO, delivering faster, reference-free preference optimization for better model alignment.

orchestra-research/ai-research-skills

This skill orchestrates distributed training with Ray Train to scale PyTorch, TF, and HuggingFace across clusters, boosting efficiency and fault tolerance.

orchestra-research/ai-research-skills

This skill helps you quantize large language models to 8-bit or 4-bit with minimal accuracy loss to reduce memory and speed up inference.

orchestra-research/ai-research-skills

This skill guarantees valid JSON/XML/code structures during generation using constrained token sampling and Pydantic-backed outputs.

orchestra-research/ai-research-skills

This skill helps you instrument, trace, evaluate, and monitor LLM applications with Phoenix for debugging, testing, and real-time observability.

knowledge-distillation

orchestra-research/ai-research-skills

This skill helps you compress large language models via knowledge distillation, preserving performance while enabling smaller, faster open-source deployments.

brainstorming-research-ideas

orchestra-research/ai-research-skills

This skill guides researchers through structured ideation frameworks to uncover high-impact research directions, offering actionable prompts and evaluation

orchestra-research/ai-research-skills

This skill provides expert guidance for fine-tuning LLaMA models with Llama-Factory, covering APIs, setup, and best practices for multimodal, 8-bit QLoRA

orchestra-research/ai-research-skills

This skill helps you apply causal interventions to PyTorch models with pyvene, enabling tracing, patching, IIT, and hypothesis testing.

transformer-lens

orchestra-research/ai-research-skills

This skill helps researchers inspect and patch transformer internals using HookPoints and activation caching to study circuits and attention.

orchestra-research/ai-research-skills

This skill helps you scale ML data processing with Ray Data, enabling streaming, GPU-accelerated transforms, and seamless integration with ML frameworks.

orchestra-research/ai-research-skills

This skill speeds high-performance RLHF training for large models with Ray and vLLM acceleration, simplifying distributed PPO GRPO DPO workflows

trl-fine-tuning

orchestra-research/ai-research-skills

This skill guides fine-tuning LLMs with TRL for instruction tuning, preference alignment, and reward-based optimization, aligning models to human feedback.

orchestra-research/ai-research-skills

This skill optimizes LLM data curation with GPU-accelerated, multi-modal cleaning, deduplication, and PII redaction to improve training data quality.

orchestra-research/ai-research-skills

This skill simplifies distributed training with HuggingFace Accelerate, enabling seamless multi-GPU/TPU setups via a four-line integration.

bigcode-evaluation-harness

orchestra-research/ai-research-skills

This skill benchmarks code generation models across 15+ tasks, providing pass@k metrics and multi-language evaluation for robust code quality.

orchestra-research/ai-research-skills

This skill speeds up structured generation and agent workflows with prefix caching for JSON outputs and tool calls, boosting latency and throughput.

weights-and-biases

orchestra-research/ai-research-skills

This skill helps you track ML experiments, visualize training, sweep hyperparameters, and manage models using Weights & Biases for streamlined MLOps.

orchestra-research/ai-research-skills

This skill helps you design, deploy, and monitor autonomous AI agents with a visual builder and persistent execution.

orchestra-research/ai-research-skills

This skill helps you orchestrate teams of autonomous agents for complex tasks with memory, roles, and production-ready workflows.

orchestra-research/ai-research-skills

This skill helps you build powerful RAG applications by ingesting documents, indexing data, and querying with LlamaIndex.

orchestra-research/ai-research-skills

This skill helps you implement open-source embedding storage and semantic search for AI apps with RAG workflows using Chroma.

orchestra-research/ai-research-skills

This skill helps you manage production-grade vector search with Pinecone, delivering low-latency, serverless indexing and hybrid search capabilities.

orchestra-research/ai-research-skills

This skill helps you implement high-performance vector similarity search for production RAG systems with fast nearest neighbor and flexible filtering.

orchestra-research/ai-research-skills

This skill extracts and validates structured data from LLM responses using Pydantic, with automatic retries and real-time streaming.

orchestra-research/ai-research-skills

This skill helps you debug, evaluate, and monitor LLM applications with LangSmith observability, capturing traces, datasets, and metrics for reliable AI

orchestra-research/ai-research-skills

This skill helps you generate music or sounds from text descriptions using AudioCraft, enabling melody-conditioned and stereo audio output.

orchestra-research/ai-research-skills

This skill helps you perform vision-language tasks such as captioning, VQA, and multimodal chat using BLIP-2 with frozen encoders.

orchestra-research/ai-research-skills

This skill helps you build and use vision-language AI chat capabilities, enabling multi-turn image analysis and VQA with LLaVA.

segment-anything

orchestra-research/ai-research-skills

This skill enables zero-shot image segmentation with the Segment Anything Model, allowing prompts by points, boxes, or masks and automatic mask generation.

orchestra-research/ai-research-skills

This skill helps you train large-scale Mixture of Experts models with DeepSpeed or HuggingFace efficiently, reducing compute while expanding capacity.

orchestra-research/ai-research-skills

This skill enables robust multilingual speech-to-text and translation workflows using Whisper across podcasts, meetings, and noisy audio for accurate

orchestra-research/ai-research-skills

This skill helps extend transformer context windows for long documents using RoPE, YaRN, ALiBi, and position interpolation to improve efficiency and

creative-thinking-for-research

orchestra-research/ai-research-skills

This skill helps researchers generate genuinely novel CS and AI ideas by applying cognitive science frameworks like combinatorial creativity and constraint