3. Data, Science & AI

Connects AI agents to external platforms (HubSpot, Salesforce) using Nango with header authentication for seamless data integration.

Configures models, tools, and options for the kenken workflow, enabling AI/ML pipeline setup and customization.

Enables seamless task transfer between AI agents in single-agent orchestration mode, maintaining workflow continuity.

Structures MCP tool calls and interprets results to fulfill user requests in AI agent systems.

Analyzes Ada's AI chatbot CSAT data to provide actionable recommendations for improving customer satisfaction and addressing dissatisfaction.

Automates large context processing, media analysis, web browsing, and code refactoring via Gemini CLI with token usage tracking.

Creates specialized Claude Code skills with optimized model configurations based on task analysis, enabling tailored AI workflows and tool integrations.

Intelligent system mapping user tasks to optimal analysis tools using hybrid retrieval techniques.

Enables a single Human-in-the-Loop (HITL) iteration with full visibility for efficient learning before stepping away.

AI engine analyzing Gulf market trends from TikTok, Meta, Amazon.ae, and Google Trends to identify high-demand, low-competition product opportunities with interactive dashboards.

Automates GPU performance optimization for AI models through benchmarking, agent coordination, and report generation with a single command.

Runs sequential GPU benchmarks across optimization worktrees to evaluate performance without interference, post-optimization.

Reads, analyzes, and generates Excel spreadsheets (.xlsx, .xlsm, .csv, .tsv) for data extraction and manipulation tasks.

Assists in understanding emergent permutations within the Universal Axiom intelligence framework, specifically how multiplicative formulas generate novel insights through dynamic variable interactions.

Analyzes and combines successful optimization agents to maximize system speedup after benchmark validation.

Fetches data from Notion databases and pages, including querying rows, retrieving page content, and database schemas.

Generates custom SVG charts from HTML data points for data visualization.

Generates comprehensive markdown reports by aggregating benchmark results, anti-triche analysis, and agent summaries from completed optimization phases.

Orchestrates persistent memory, task planning, and state management for AI agent systems to enable complex multi-step workflows and context retention.

Expert analysis of longitudinal sequence data using TraMineR (R) and TanaT (Python) for state/event sequences, distance metrics, clustering, and visualization.

Specializes in generative AI model development, LLM integration, and MLOps for deploying and managing AI systems.

Automates context window reset for Ralph agents during long tasks to maintain LLM performance and prevent context overflow.

Provides advanced predictive modeling, statistical analysis, and data-driven insight discovery for complex datasets.

Provides LLM-powered workflows for exploring and analyzing data within dbt projects, enhancing data exploration with AI observability.