Skills

Queries ChEMBL's extensive bioactive molecule database to retrieve compound, target, and bioactivity data for drug discovery research.

documentation review skill from ovachiever/droid-tings

Parses FCS flow cytometry files, extracting event data as NumPy arrays, reading metadata, and converting to CSV/DataFrame for scientific data preprocessing.

Guides fine-tuning language models using Group Relative Policy Optimization (GRPO) for structured reasoning and task-specific training with TRL.

Extracts and validates structured data from LLM responses using Pydantic schemas, supporting multiple providers with automatic retries and type safety.

Queries KEGG databases via REST API to retrieve comprehensive biological pathway, gene, compound, and molecular interaction information for academic research.

Reads, writes, and manipulates DICOM medical imaging files, enabling pixel data extraction, metadata processing, and image format conversion.

Enables world-class ML engineering by productionizing AI/ML models, implementing MLOps, and building scalable machine learning infrastructure with cutting-edge tools and practices.

Enables fast, structured LLM generation with RadixAttention prefix caching for JSON, regex, and agentic workflows with 5× faster inference.

Visualizes and tracks machine learning training metrics, model architectures, and performance across PyTorch and TensorFlow experiments.

Performs fast, nonlinear dimensionality reduction for high-dimensional data visualization, clustering preprocessing, and machine learning embeddings using UMAP algorithm.

Provides comprehensive LaTeX templates and formatting guidelines for scientific journals, conferences, research posters, and grant proposals across multiple disciplines.