3. Data, Science & AI

Provides implementation guidance for Vercel AI SDK v5, covering chat interfaces, streaming, tool calling, and text embeddings for AI application development.

Extracts clean, readable text from web articles, blog posts, and documentation by removing ads and clutter.

Enables execution, conversion, and exploration of Jupyter notebooks within the wc_simd project.

Manages database operations, migrations, Row-Level Security (RLS) policies, and performance tuning for efficient data handling.

Generates text embeddings using Qwen3 models through HuggingFace TEI for NLP tasks, supporting batch processing and service configuration.

Manages persistent memory across Claude conversations to prevent repetition of mistakes and apply learned patterns from previous interactions.

Systematically elicits necessary context for complex problems using adaptive prompt patterns to clarify requirements and verify understanding in AI interactions.

Automates CSV/XLSX data loading into DuckDB with validation and generates an ETL markdown report.

Web-based SPA for statistical analysis with outlier detection (Z-Score, IQR) and hypothesis testing (T-test, ANOVA), processing data locally via CSV/Recharts visualization.

Analyzes Pingo Doce purchase data against nutritional guidelines to identify dietary gaps and provide personalized health recommendations.

Optimizes and designs effective prompts for LLM interactions, improving output quality and reliability in AI applications.

Analyzes web traffic and measurement data using Python, pandas, and SQL for statistical insights and visualization.

Reconciles normalized CI, PL, and BL fields across datasets and outputs a mismatch report for data validation.

Enables on-device AI implementation using Apple's Foundation Models for summarization, extraction, and classification with type-safe structured output via @Generable.

Simplifies SQLite database operations including persistence, queries, writes, and migrations via the SQLiteData library.

Provides tools for creating, editing, and building Observable Notebooks with database queries, data loaders, and static site generation for data science.

Tracks blog post metrics and generates actionable insights to improve content performance and engagement.

Analyzes Oura ring biometric data (HRV, sleep quality, readiness scores) to provide personalized scheduling, task prioritization, and recovery recommendations for enhanced productivity.

Enables querying, schema exploration, and data distribution analysis for Hive tables in the Wellcome Collection data warehouse.

Generates markdown report for HS risk scoring with data flags (format, missing, haz, oog).

Processes aviation fuel surcharge bills from Excel files by detecting headers, matching columns, parsing dates/routes, and fetching contract numbers via API.

Manages Vision-Language Model (VLM) embedding services, including starting Flask-based services, running embedding jobs, and training AE3D autoencoders.

Processes IIIF manifests from the Wellcome Collection API, enabling metadata extraction and canvas processing for digital image collections.

Manages Ollama AI models for local deployment, enabling model pulling, listing, preset switching, and disk usage checks.