orchestra-research/ai-research-skills

1

speculative-decoding Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use wh…

orchestra-research/ai-research-skills

750

2

training-llms-megatron Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Us…

orchestra-research/ai-research-skills

750

3

distributed-llm-pretraining-torchtitan Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use whe…

orchestra-research/ai-research-skills

749

4

nanogpt Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learni…

orchestra-research/ai-research-skills

749

5

outlines Guarantee valid JSON/XML/code structure during generation, use Pydantic models for type-safe outputs, support local m…

orchestra-research/ai-research-skills

749

6

pyvene-interventions Provides guidance for performing causal interventions on PyTorch models using pyvene's declarative intervention frame…

orchestra-research/ai-research-skills

749

7

clip OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross…

orchestra-research/ai-research-skills

748

8

llama-factory Expert guidance for fine-tuning LLMs with LLaMA-Factory - WebUI no-code, 100+ models, 2/3/4/5/6/8-bit QLoRA, multimod…

orchestra-research/ai-research-skills

748

9

llamaindex Data framework for building LLM applications with RAG. Specializes in document ingestion (300+ connectors), indexing,…

orchestra-research/ai-research-skills

748

10

sparse-autoencoder-training Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network act…

orchestra-research/ai-research-skills

748

11

gptq Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on c…

orchestra-research/ai-research-skills

747

12

nemo-curator GPU-accelerated data curation for LLM training. Supports text/image/video/audio. Features fuzzy deduplication (16× fa…

orchestra-research/ai-research-skills

747

13

simpo-training Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 poi…

orchestra-research/ai-research-skills

747

14

slime-rl-training Provides guidance for LLM post-training with RL using slime, a Megatron+SGLang framework. Use when training GLM model…

orchestra-research/ai-research-skills

747

15

transformer-lens-interpretability Provides guidance for mechanistic interpretability research using TransformerLens to inspect and manipulate transform…

orchestra-research/ai-research-skills

747

16

guidance Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and…

orchestra-research/ai-research-skills

746

17

llamaguard Meta's 7-8B specialized moderation model for LLM input/output filtering. 6 safety categories - violence/hate, sexual …

orchestra-research/ai-research-skills

746

18

model-merging Merge multiple fine-tuned models using mergekit to combine capabilities without retraining. Use when creating special…

orchestra-research/ai-research-skills

746

19

nemo-guardrails NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-c…

orchestra-research/ai-research-skills

745

20

evaluating-code-models Evaluates code generation models across HumanEval, MBPP, MultiPL-E, and 15+ benchmarks with pass@k metrics. Use when …

orchestra-research/ai-research-skills

744

21

torchforge-rl-training Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Us…

orchestra-research/ai-research-skills

744

22

autogpt-agents Autonomous AI agent platform for building and deploying continuous agents. Use when creating visual workflow agents, …

orchestra-research/ai-research-skills

743

23

constitutional-ai Anthropic's method for training harmless AI through self-improvement. Two-phase approach - supervised learning with s…

orchestra-research/ai-research-skills

743

24

implementing-llms-litgpt Implements and trains LLMs using Lightning AI's LitGPT with 20+ pretrained architectures (Llama, Gemma, Phi, Qwen, Mi…

orchestra-research/ai-research-skills

743

25

langsmith-observability LLM observability platform for tracing, evaluation, and monitoring. Use when debugging LLM applications, evaluating m…

orchestra-research/ai-research-skills

743

26

miles-rl-training Provides guidance for enterprise-grade RL training using miles, a production-ready fork of slime. Use when training l…

orchestra-research/ai-research-skills

743

27

nnsight-remote-interpretability Provides guidance for interpreting and manipulating neural network internals using nnsight with optional NDIF remote …

orchestra-research/ai-research-skills

743

28

skypilot-multi-cloud-orchestration Multi-cloud orchestration for ML workloads with automatic cost optimization. Use when you need to run training or bat…

orchestra-research/ai-research-skills

743

29

mamba-architecture State-space model with O(n) complexity vs Transformers' O(n²). 5× faster inference, million-token sequences, no KV ca…

orchestra-research/ai-research-skills

742

30

model-pruning Reduce LLM size and accelerate inference using pruning techniques like Wanda and SparseGPT. Use when compressing mode…

orchestra-research/ai-research-skills

742

31

rwkv-architecture RNN+Transformer hybrid with O(n) inference. Linear time, infinite context, no KV cache. Train like GPT (parallel), in…

orchestra-research/ai-research-skills

742

32

nemo-evaluator-sdk Evaluates LLMs across 100+ benchmarks from 18+ harnesses (MMLU, HumanEval, GSM8K, safety, VLM) with multi-backend exe…

orchestra-research/ai-research-skills

741

33

pinecone Managed vector database for production AI applications. Fully managed, auto-scaling, with hybrid search (dense + spar…

orchestra-research/ai-research-skills

741

34

blip-2-vision-language Vision-language pre-training framework bridging frozen image encoders and LLMs. Use when you need image captioning, v…

orchestra-research/ai-research-skills

740

35

phoenix-observability Open-source AI observability platform for LLM tracing, evaluation, and monitoring. Use when debugging LLM application…

orchestra-research/ai-research-skills

740

36

lambda-labs-gpu-cloud Reserved and on-demand GPU cloud instances for ML training and inference. Use when you need dedicated GPU instances w…

orchestra-research/ai-research-skills

739

37

llava Large Language and Vision Assistant. Enables visual instruction tuning and image-based conversations. Combines CLIP v…

orchestra-research/ai-research-skills

739

38

experiment-tracking-swanlab Provides guidance for experiment tracking with SwanLab. Use when you need open-source run tracking, local or self-hos…

orchestra-research/ai-research-skills

688

39

systems-paper-writing Comprehensive guide for writing systems papers targeting OSDI, SOSP, ASPLOS, NSDI, and EuroSys. Provides paragraph-le…

orchestra-research/ai-research-skills

684

40

presenting-conference-talks Generates conference presentation slides (Beamer LaTeX PDF and editable PPTX) from a compiled paper with speaker note…

orchestra-research/ai-research-skills

674

41

ml-training-recipes Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug disco…

orchestra-research/ai-research-skills

663

42

fine-tuning-openvla-oft Fine-tunes and evaluates OpenVLA-OFT and OpenVLA-OFT+ policies for robot action generation with continuous action hea…

orchestra-research/ai-research-skills

659

43

fine-tuning-serving-openpi Fine-tune and serve Physical Intelligence OpenPI models (pi0, pi0-fast, pi0.5) using JAX or PyTorch backends for robo…

orchestra-research/ai-research-skills

657

44

evaluating-cosmos-policy Evaluates NVIDIA Cosmos Policy on LIBERO and RoboCasa simulation environments. Use when setting up cosmos-policy for …

orchestra-research/ai-research-skills

656

45

evolving-ai-agents Provides guidance for automatically evolving and optimizing AI agents across any domain using LLM-driven evolution al…

orchestra-research/ai-research-skills

625

46

ara-rigor-reviewer Performs ARA Seal Level 2 semantic epistemic review on Agent-Native Research Artifacts, scoring six dimensions (evide…

orchestra-research/ai-research-skills

588

47

ara-compiler Compiles any research input — PDF papers, GitHub repositories, experiment logs, code directories, or raw notes — into…

orchestra-research/ai-research-skills

577

48

ara-research-manager Records research provenance as a post-task epilogue, scanning conversation history at the end of a coding or research…

orchestra-research/ai-research-skills

575

49

writing-systems-papers Paragraph-level structural blueprint for 12-page systems papers targeting OSDI, SOSP, ASPLOS, NSDI, and EuroSys. Prov…

orchestra-research/ai-research-skills

5

50

pytorch-fsdp Expert guidance for Fully Sharded Data Parallel training with PyTorch FSDP - parameter sharding, mixed precision, CPU…

orchestra-research/ai-research-skills

2