3. Data, Science & AI

Analyzes YouTube video transcripts through download, chunking, and AI-driven context-aware insights for research and project alignment.

A graphical user interface for interacting with Vector RAG systems, enabling intuitive data retrieval and generation workflows.

Extracts structured text, tables, and images from PDF documents using advanced libraries for data analysis and reporting.

Creates, optimizes, and analyzes prompts for LLMs including model selection guidance for Claude, GPT, and Gemini.

Transforms vague user requests into structured, high-quality AI prompts using XML, examples, and chain-of-thought reasoning with transparent preview.

Automates aggregation of flat building elements into hierarchical construction documents with standardized 4-level cost breakdown structure.

Converts construction estimate documents (Excel, CSV, PDF) from Supabase into EstimationElement format for structured data analysis.

Enables programmatic access to OpenAI's API for AI model interactions including text generation, embeddings, and language processing.

Analyzes crypto market data to compute rolling correlations, betas vs BTC, and VWAP status for rotation eligibility checks.

Summarizes ETF net flows and derivatives context (OI Δ, funding, basis, skew, liquidation bands) from local JSON/CSV files for market analysis at open/close or leverage reset events.

Collects and analyzes GitHub PRs with AI review comments using GitHub CLI and jq, organizing data into structured PR metadata.

Compares AI-generated cost estimates with actual company quotes from PDFs to identify pricing discrepancies and new items for validation.

Automates end-to-end processing of D&D session videos, including transcription and knowledge extraction for structured data.

Troubleshoots FFmpeg integration errors, audio extraction failures, and video codec issues in media processing workflows.

Analyzes D&D session data to extract and summarize key campaign elements including NPCs, locations, quests, items, and factions for narrative review and entity discovery.

Python framework for rapidly building interactive data visualizations, dashboards, and ML demos without front-end coding.

High-performance DataFrame library for fast data manipulation, analysis, and ETL with lazy evaluation on large datasets.

Provides fast in-process SQL analytics on data files (CSV, Parquet, JSON) and DataFrames (pandas, Polars) without memory loading.

Optimizes RAG performance with reranking, caching, parallel processing, and production deployment patterns for improved retrieval quality and scalability.

Evaluates RAG system performance using hit rate, MRR, and faithfulness metrics to compare retrieval strategies and embeddings.

Sets up RAG pipelines using LlamaIndex for document chunking, embedding generation, and vector/hybrid retrieval.

Provides integration patterns for deepagents framework to build AI agents, handle planning, filesystem operations, and orchestrate subagents using LangGraph.

Automates data processing pipelines for web-scraped data, including validation, cleaning, and storage to CSV, JSON, or databases.

Processes and generates multimedia content (audio, image, video, documents) using Google Gemini API and specialized AI models for multimodal applications.