3. Data, Science & AI

Creates and manages RDF ontologies using Turtle syntax for semantic data modeling, domain specifications, and knowledge graph architecture.

Provides parallel orchestration for 10 agents through an atomic cycle (fan-out → construction → collision → convergence) in AI systems.

Analyzes images using Gemini Flash vision model via REST API for captioning, OCR, and multi-image comparison.

Provides institutional-grade data cleaning, tick physics auditing, and gap reconstruction for financial market data.

AI-powered semantic search and document parsing for finding conceptually related content in technical documentation and codebases.

Automates skill enhancement in Univrs.io by observing execution and integrating critique feedback to generate improved versions.

Generates structured guidelines for effective AI prompts and coding agent rules to optimize model interactions and task execution.

Coordinates multiple AI agent command-line interfaces to execute complex tasks, gather alternative approaches, and synthesize diverse opinions.

Provides access to US Census Bureau data including demographics, income, housing, and population statistics via Census API queries.

Provides access to Cook County government data portals including property assessments, court cases, and medical examiner records through datacatalog.cookcountyil.gov.

Provides access to Chicago city datasets including crime, permits, and 311 data through the city's open data portal.

Analyzes uploaded images to extract style elements and generate structured JSON profiles for AI image generation prompts and style recreation.

Creates specialized Claude Code sub-agents with customizable system prompts and tool configurations for task-specific AI workflows.

Matches community members with relevant expertise using AI-powered vector search for semantic similarity.

Context for ground station package handling telemetry reception, mission control interface, and real-time flight monitoring systems.

Extracts transcripts, generates summaries, and creates Q&A highlights from YouTube videos or local media files using AI-powered analysis.

Analyzes meeting transcripts using AI to identify communication patterns, including conflict avoidance, filler words, and listening gaps for professional development.

Accelerates LLM fine-tuning with Unsloth, reducing memory usage by 80% and doubling speed for models including Llama, Mistral, and Gemma.

SuperBPE provides efficient tokenization for LLMs, reducing token count by 20-33% to lower costs and expand context windows. Supports training, validation, and integration.

Analyzes and compares tokenizers for efficiency, enabling integration with Unsloth models in AI/ML workflows.

Provides context and utilities for camera capture, video processing, audio recording, and mission logic in flight computer systems.

Provides advanced techniques for optimizing LLM fine-tuning, including hyperparameter tuning and monitoring for peak model performance.

Deploys fine-tuned machine learning models to production using vLLM, HuggingFace Hub, and Docker for efficient serving.

Manages a points-based reward system for MLAI platform users, enabling balance checks, coworking bookings, task claims, and reward redemptions.