3. Data, Science & AI

Processes DOCX exam papers into structured markdown with image analysis, context matching, and YAML frontmatter for educational content.

Automates extraction of toxicity and peptide experimental data from open-access biomedical literature, including full text and supplementary materials, for downstream analysis.

Monitors AI capabilities in production, tracking value realization, usage patterns, risk indicators, and generating stakeholder-specific reports.

Provides mathematical utilities for LÖVE game development, including random number generation and geometric calculations.

Audits AI systems for fairness, generating bias assessments, metrics, mitigation strategies, and documentation to ensure ethical deployment.

Designs data annotation workflows including guidelines, quality control, and labeler training to prepare high-quality datasets for machine learning model training.

Assesses codebase maturity across eight technical pillars to evaluate readiness for autonomous AI development and identify gaps.

Expert prompt engineering for optimizing system prompts, user prompts, and few-shot examples to enhance LLM performance across Claude, GPT, and other models.

Manages production AI/ML model lifecycle with registry entries, drift detection, retraining triggers, and version control.

Collaboratively crafts high-quality prompts for AI models through iterative dialogue, optimizing system instructions and AI behavior.

Generates scaling plans, readiness assessments, and rollout strategies for transitioning successful AI pilots to enterprise-scale production deployments.

Guides data analysis by clarifying end goals first, then applying Python and pandas to extract actionable insights.

Guides creation, configuration, and usage of Claude Code subagents and the Task tool for specialized AI agent workflows.

Decomposes large specifications into structured agent skills for efficient AI navigation, enabling seamless knowledge access in complex documentation and frameworks.

Automates the import, categorization, and Cooklang format validation of recipes from web URLs.

Generates optimized prompts for multi-stage Claude-to-Claude pipelines enabling research, planning, and execution stages in AI workflows.

Generates new AI agent skills following the Agent Skills specification, enabling users to define capabilities and document AI agent patterns.

Evaluates AI capability test results against acceptance criteria to determine production deployment readiness, providing structured go/no-go verdicts with gap analysis.

Indexes research projects and documents, extracts metadata, and enables semantic search via local RAG for portfolio management.

Enables structured hypothesis testing and reproducible experimental design for data science and AI model validation.

Provides physics-based calculations for Small-Angle X-ray Scattering (SAXS) data analysis, including Rg, I(0), and Crystallinity Index.

Audits machine learning models for demographic bias using the Fairlearn library, ensuring equitable outcomes in predictive systems.

Constructs precise URL queries for the US Department of Education College Scorecard API to fetch and clean institutional data.

Enables SQL querying of JGI's GOLD and IMG genomics databases within the Dremio Lakehouse for genomic data exploration.