3. Data, Science & AI

Optimizes LLM agent prompts for clarity, structure, and behavior. Enhances prompt engineering and manages LangGraph agent configurations.

Guides creation and maintenance of custom agents for VS Code Copilot Chat, covering intent discovery, guardrail writing, and tool selection.

Reviews, fixes, and guides creation of Builder.io Fusion rules files to maximize AI performance and address ignored rules.

Organizes experiment logs, results, and metadata for Python-based research code, enhancing reproducibility and analysis.

Systematically analyzes datasets to identify structure, quality issues, distributions, and correlations through automated profiling and visualization.

Generates publication-quality data visualizations including chart selection, dashboard design, and interactive elements following data science best practices.

Builds data-driven narratives for presenting analysis results and communicating insights through storytelling frameworks in reports and presentations.

Converts data analysis results into actionable business insights, connecting findings to strategic impact for stakeholder communication.

Compresses memory data into structured summaries using low-cost AI models, triggered by count or token thresholds.

Efficiently packages context documents and structures prompts for AI analysis with Claude, optimizing complex task preparation.

Validates AI-generated output fields for correctness and completeness, auto-triggered by AI-related keywords and truncation issues.

Provides intelligent memory search using semantic (vector), keyword, and recent query methods with automatic fallback and optional summaries.

Maps database table relationships and generates ERD documentation for schema exploration and understanding.

Monitors AI context token usage, warns at 30k, 50k, and 70k thresholds, and provides context management recommendations.

Monitors GitHub's trending AI projects, analyzing activity metrics like stars and commits to provide insights on the most active and promising open-source AI initiatives.

Extracts text, tables, and content from PDFs using OCR and document analysis for data processing and integration.

Automates verification by dispatching multiple independent agents, aggregating results via consensus ratio, and cross-checking unique findings to ensure accuracy.

Expert guidance for SQLMesh projects: creating data models, reviewing deployment plans, validating changes, debugging, and production deployment.

Provides production-ready integration for OpenAI and Azure OpenAI APIs, covering structured outputs, function calling, cost optimization, and error handling.

Real-time fraud detection system using machine learning, behavioral biometrics, and device fingerprinting for African fintech transaction monitoring.

Expert assistance for PostgreSQL database tasks including query writing, optimization, schema design, and SQL validation in code.

Analyzes Excel spreadsheets using formulas, charts, and pivot tables to extract insights from structured data.

Generates comprehensive CSV analysis including summary statistics, pattern identification, and visualizations with automatic adaptation to data types like sales and financial.

Configures axis mappings and display settings for charts across all chart types in Drizzle Cube, enabling tailored data visualization.