63 skills · sorted by relevance
Design scheduled AI workflows for document ingestion, batch processing, and model training using Prefect, Airflow, or Dagster. Compare orchestration tools and plan task dependencies, retries, caching, and monitoring through implementation fragments.
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.
The English summary is being prepared. Showing the original description.
Deploy machine learning models to production serving infrastructure using MLflow, BentoML, or Seldon Core with REST/gRPC endpoints, implement autoscaling, monitoring, and A/B testing capabilities for high-performance model inference at scale. Use when deploying trained models for real-time inference, setting up REST or gRPC prediction APIs, implementing autoscaling for variable load, running A/B tests between model versions, or migrating from batch to real-time inference.
The English summary is being prepared. Showing the original description.
Otimiza inferência de LLM com NVIDIA TensorRT para máxima vazão e latência mínima. Use para implantação em produção em GPUs NVIDIA (A100/H100), quando você precisa de inferência 10-100x mais rápida que PyTorch, ou para servir modelos com quantização (FP8/INT4), batching em voo e escalabilidade multi-GPU.
The English summary is being prepared. Showing the original description.
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.
The English summary is being prepared. Showing the original description.
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
The English summary is being prepared. Showing the original description.
Run diffusers generation pipelines from a terminal and save their outputs. Inspect accepted inputs without downloading model weights, apply LoRA adapters during generation, run remotely through HF Jobs, and inspect software and GPU environment details.
Plan, create, and configure production-ready Google Kubernetes Engine (GKE) clusters using the golden path Autopilot configuration. Covers Day-0 checklist, Autopilot vs Standard, networking (private clusters, VPC-native, Gateway API), security (Workload Identity, Secret Manager, RBAC hardening), observability, scaling, cost optimization, and AI/ML inference. WHEN: create GKE cluster, provision GKE environment, design GKE networking, secure GKE, optimize GKE cost, GKE autoscaling, GKE inference, GKE upgrade, GKE observability, GKE multi-tenancy, GKE batch, GKE HPC, GKE compute class.
The English summary is being prepared. Showing the original description.
Serve LLMs com alta throughput usando PagedAttention do vLLM e continuous batching. Use ao fazer deploy de APIs LLM em produção, otimizar latência/throughput de inferência, ou servir modelos com memória GPU limitada. Suporta endpoints compatíveis com OpenAI, quantização (GPTQ/AWQ/FP8) e tensor parallelism.
The English summary is being prepared. Showing the original description.
Choose OpenRouter models and routers, query live catalog data, use structured outputs and batch inference, and debug routing with metadata. Use when working with OpenRouter APIs, openrouter/auto, openrouter/free, model discovery, pricing, providers, batch jobs, or structured JSON responses.
The English summary is being prepared. Showing the original description.
Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.
The English summary is being prepared. Showing the original description.
Use Hugging Face Transformers to load pretrained models, generate text, and classify text, images, or audio. Prepare model inputs, control generation settings, and fine-tune models on custom datasets with the Trainer API.
Evaluate and use SSH Remote Compute before choosing where to run GPU, high-memory, parallel, batch, model-inference, bioinformatics, or other long-running scientific work; supports short remote commands and asynchronous jobs with automatic harvest and analysis.
The English summary is being prepared. Showing the original description.
Pull pre-computed ortholog calls from public databases (OrthoDB, Ensembl Compara, OMA browser, eggNOG, PANTHER, KEGG Orthology, HomoloGene) via their REST APIs. Use when orthologs are already curated upstream, when the question is "what is the X ortholog of Y" rather than "how to infer orthology de novo", when batch-mapping gene IDs across species, or when comparing the resources for consensus calls. Encodes confidence-level semantics, 1:1 vs 1:many vs many:many, HomoloGene deprecation, and when to defect to de novo computation.
The English summary is being prepared. Showing the original description.
Deploy and manage vLLM for high-throughput LLM inference. Configure continuous batching, tensor parallelism, quantization, and OpenAI-compatible API endpoints for production LLM serving.
The English summary is being prepared. Showing the original description.
Orchestrate AI/ML pipelines for data ingestion, model training, batch inference, and RAG indexing using Prefect, Airflow, or Dagster. Build reliable, observable, and retriable workflows for production AI systems.
The English summary is being prepared. Showing the original description.
Decision SOP for serving LLMs with vLLM. Covers PagedAttention mental model, quantization/parallelism/batching tradeoffs, OOM triage, and when NOT to use vLLM. Activates when a coder-agent is choosing or tuning an inference engine, debugging vLLM throughput/latency/OOM, or comparing vLLM against TGI/SGLang/TensorRT-LLM/llama.cpp.
The English summary is being prepared. Showing the original description.
Orchestrate AI/ML pipelines for data ingestion, model training, batch inference, and RAG indexing using Prefect, Airflow, or Dagster.
The English summary is being prepared. Showing the original description.
| Type-safe ORM for Cloudflare D1 databases using Drizzle. Use when: building D1 database schemas, writing type-safe SQL queries, managing migrations with Drizzle Kit, defining table relations, implementing prepared statements, using D1 batch API, or encountering D1_ERROR, transaction errors, foreign key constraint failures, or schema inference issues.
The English summary is being prepared. Showing the original description.
Pull pre-computed ortholog calls from public databases (OrthoDB, Ensembl Compara, OMA browser, eggNOG, PANTHER, KEGG Orthology, HomoloGene) via their REST APIs. Use when orthologs are already curated upstream, when the question is "what is the X ortholog of Y" rather than "how to infer orthology de novo", when batch-mapping gene IDs across species, or when comparing the resources for consensus calls. Encodes confidence-level semantics, 1:1 vs 1:many vs many:many, HomoloGene deprecation, and when to defect to de novo computation.
The English summary is being prepared. Showing the original description.
LLM serving for latency, batching, caching, quantization, and routing. Use when tuning vLLM or SGLang, sizing KV cache, adding speculative decoding, or cutting serving cost.
The English summary is being prepared. Showing the original description.
Design and configure model serving infrastructure — endpoint scaling, batching, GPU allocation. Use when asked to "serve this model", "design an inference endpoint", or "size GPU allocation for serving".
The English summary is being prepared. Showing the original description.
Design cost reduction strategies — model tiering, prompt compression, caching, batch inference. Use when asked to "reduce our AI costs", "set up model tiering", or "cut LLM spend".
The English summary is being prepared. Showing the original description.
MLOps and the production ML lifecycle -- model packaging and serving, CI/CD for ML, experiment tracking, model registries, reproducibility, production monitoring for data and concept drift, retraining pipelines, A/B and shadow deployment, and rollback. Covers batch vs online/real-time inference, REST endpoints, feature stores, data and version pinning, deterministic pipelines, performance-decay detection, and retraining triggers. Use when deploying models to production, serving predictions, monitoring for data or concept drift, setting up ML CI/CD, tracking experiments, managing a model registry, or planning retraining, shadow rollout, and rollback.
The English summary is being prepared. Showing the original description.