3. Data, Science & AI

Manages Langfuse prompts with version control, deployment labels, and comparison for LLM application development.

Builds foundational evaluation infrastructure for AI/ML models, including datasets, graders, harness, and baselines required before optimization.

Configures agent evaluation pipelines including flow discovery, quality dimensions, dataset creation, and judge prompts for AI agent testing.

Diagnoses root causes in LLM application workflows by combining Langfuse trace data with codebase investigation to resolve failures and improve output quality.

Runs Langfuse experiments to evaluate LLM performance, compare model and prompt variations, and analyze failures.

Executes and analyzes LLM evaluation experiments using Langfuse, including prompt testing, dataset evaluation, and LLM-as-judge comparisons.

Analyzes score trends, regressions, and quality metrics distributions over time for performance evaluation and improvement.

Surgically retrieves Langfuse observability data for debugging and analyzing LLM application traces with multiple output modes.

Defines goal, constraints, and adjustable parameters for optimization problems in AI and data science frameworks.

Enables session-level analysis in Langfuse to inspect, debug multi-turn conversations, identify errors, and evaluate performance metrics.

Generates ASCII-based data visualizations including charts, graphs, and progress bars for terminal display in Claude Code responses.

Configures Langfuse tracing, observability, and scoring for Python-based LLM pipelines to monitor and debug AI model interactions.

Enables cross-platform event tracking for user interactions and feature usage with analytics providers like Firebase and Adjust.

Manages Langfuse datasets for LLM experiment validation and regression testing, including dataset creation, trace curation, and test set building.

Generates audio files from text segments using MiniMax TTS API with built-in error handling and retry logic.

Manages Langfuse datasets for LLM applications, enabling creation, curation of regression/golden sets, and dataset item inspection.

Initializes Serena MCP for AI-enhanced semantic analysis of code, enabling deeper code structure and meaning interpretation during analysis.

Provides strategic guidance for evaluating, improving AI agents, selecting metrics, building datasets, and setting up iteration loops with Langfuse.

Optimizes database performance through advanced SQL query tuning, indexing strategies, and EXPLAIN plan analysis.

Retrieves and analyzes Langfuse traces, runs, and metadata for debugging and optimizing LLM application workflows.

Analyzes LLM model quality metrics including score trends, regressions, and distributions across releases and environments in Langfuse.

Specializes in relational database design, schema modeling, normalization, indexing, and migration planning for optimal data integrity and performance.

Provides a hypothesis-driven methodology for systematically improving AI agent performance through iterative optimization cycles.

Configures and manages Cursor IDE agents for parallel AI-assisted coding, supporting background, cloud, and multi-agent development workflows.