3. Data, Science & AI

Extends Convex database schema with types, indexes, and validators for new tables or structural modifications.

Accesses and retrieves company data for over 1.2 million registered businesses in Norway via the Brønnøysundregistrene registry.

Automates conversion of AI instruction files to PromptScript format for standardized LLM prompt management.

Ensures structural integrity and chronological ordering of organizational decisions in a ledger system for accurate historical tracking.

AI-powered tool that analyzes subtitles to identify optimal video clip points and segments videos using FFmpeg.

Identifies similar historical patterns in price, volume, or event data to provide analytical insights for decision-making.

Initializes and runs game simulations using coordinated multi-agent architecture for playtesting and rule validation.

Generates personalized learning content recommendations using learner analytics and profiles for adaptive learning systems and content discovery.

Configures model load balancing and fallback routing between workers in Spice for optimized AI model serving and request distribution.

Provides comprehensive tools for creating, editing, and analyzing spreadsheets with support for formulas, formatting, and data visualization.

Processes PDF documents via slicing, OpenCV-based image detection, and merging for phase 1 data ingestion tasks.

Estimates GPU memory requirements for Megatron-based MoE and dense models, enabling efficient resource planning and training configuration optimization.

Intelligently selects optimal chart types from 26 options and generates visualizations via Python script based on user data specifications.

Builds custom AI agents for diverse domains, covering architecture, capabilities, and autonomous workflows using LLM-based systems.

Integrates Scorable LLM-as-a-Judge to evaluate, monitor, and add guardrails to LLM application interactions.

Provides production-ready PostgreSQL schema patterns for social, e-commerce, and SaaS applications with Row Level Security guidance.

Manages PySpider web scraping tasks via natural language, enabling collection of social media, e-commerce, and SEO data from platforms like Reddit, Amazon, and SEMrush.

Enables RESTful API access to Google Sheets for standard CRUD operations, transforming spreadsheets into a database-like data store.

VM0 CLI enables secure sandboxed development, deployment, and execution of AI agents with project management and scheduling capabilities.

Builds data-driven dashboards to measure learning outcomes and correlate with business performance metrics.

Automates conversion of AI instruction files to PromptScript format for improved LLM prompt management and compatibility.

Designs MCP resources to structure content for LLM consumption, supporting static and dynamic resource creation in xmcp.

Real-time control of Tavus CVI conversations via Interactions Protocol, enabling text-to-speech, interruptions, context injection, and tool handling.

Analyzes CSV data to generate summary statistics, identify patterns, and create visualizations with adaptive analysis for various data types using Python and pandas.