3. Data, Science & AI

An AI-driven integrated trading platform that uses research insights to enhance trading decisions and execution in financial markets.

Diagnoses and resolves common web scraping issues in Supacrawl, including empty responses, timeouts, and anti-bot blocks.

Enables cost-effective continual learning in AI models by capturing task outcomes, scoring performance, and deriving token priors for pattern extraction without model weight changes.

Enables creation, editing, and analysis of spreadsheets with formula handling, formatting, and data visualization for data processing tasks.

Provides conversational AI capabilities for text generation, question answering, and dialogue using OpenAI's language model.

Retrieves specific vector entries by key from the Agentuity cloud platform, requiring authentication for secure data access.

High-performance Rust web crawler optimized for extracting and structuring web content into LLM-ready formats for AI/ML pipelines and site analysis.

Tracks and analyzes website and app user behavior using Google Analytics 4's data collection and reporting capabilities.

Provides temporal data processing for event tracking, gap detection, and feeding time-aware experiences to the I-layer.

Routes tasks to small AI models by default, escalating to large models only on low confidence, optimizing cost and speed without sacrificing accuracy.

Runs BEM-based acoustic simulations on horn geometries, delivering performance metrics, impedance analysis, and polar pattern visualizations.

Enables Claude to leverage NotebookLM for AI-powered document analysis, audio overviews, and knowledge synthesis from diverse sources.

Creates publication-quality visualizations for fractal horn design, including 3D geometry, acoustic plots, and simulation animations.

Queryable knowledge base for Ericsson LTE/NR RAN features, parameters, counters, and technical documents with domain-specific search capabilities.

Provides Zosia with temporal context awareness, remembering events across sessions and maintaining present-moment presence for natural, human-like interactions.

Generates fractal horn geometries using space-filling curves and Mandelbrot expansion for acoustic optimization, producing STL meshes and analysis data.

Generates images and extracts pixel data from specified coordinates for technical analysis and processing.

Optimizes RAG pipelines by enhancing search quality, reranking, query expansion, and hybrid search capabilities using vector and semantic search techniques.

Analyzes task failures to extract actionable lessons, improving system accuracy by 10% via reflection and insight curation without context collapse.

Executes PostgreSQL queries on Supabase and formats results for data retrieval and analysis.

Exposes AI Factory tools via Model Context Protocol server for seamless integration with Claude Desktop's AI workflows.

Provides real-time visualization of sleep, readiness, and activity scores from the Oura Ring wearable device.

Enables knowledge extraction from Pulse Radar using Pydantic-AI agents, RAG, and embeddings for structured data processing.

Creates interactive data visualizations with pyecharts, generates HTML reports, and applies data presentation best practices.