3. Data, Science & AI

Modifies AI agent code, adds tools, and adjusts configurations for agent behavior and MCP server integration.

Builds actionable KPI dashboards with data sourcing, visualization, and anomaly detection for business insights.

Methodology for testing AI agents with test inputs, observing outputs, and iterating for accuracy and structured results.

Defines data entities, relationships, and ownership in pre-database selection documentation for large-scale data systems.

Calls a Foundry AI agent with a user prompt and returns the generated response for application integration.

Provides Gemini AI as an oracle for planning and code review opinions, not for implementation tasks.

Automatically cleans and optimizes project data files to improve efficiency and reduce storage requirements.

Custom skill for GLM model integration, enabling tailored AI interactions and task-specific workflows.

Provides a custom implementation of the minimax algorithm for AI-driven decision making in competitive game scenarios.

Enables integration with Synth AI API using Python SDK or HTTP within OpenCode and Synth-AI TUI environments.

Manages dynamic AI context through vector databases, knowledge graphs, and intelligent memory systems to orchestrate multi-agent workflows and enterprise applications.

Optimizes cloud-native database performance using advanced SQL, OLTP/OLAP tuning, and data modeling for complex analytical workloads.

Designs and implements complete machine learning pipelines for data processing, model training, and deployment.

Builds robust backtesting systems for trading strategies, addressing look-ahead bias, survivorship bias, and transaction costs to validate algorithmic models.

Builds and deploys production machine learning systems using PyTorch, TensorFlow, and modern ML frameworks with model serving and monitoring capabilities.

Specializes in implementing vector databases for semantic search and RAG applications using Pinecone, Weaviate, and Qdrant.

Designs PostgreSQL schemas with best practices, data types, indexing, constraints, and performance patterns.

Optimizes LLM performance and output reliability through advanced prompt engineering patterns for production deployment.

Provides best practices for building and organizing data transformations with dbt, including testing, documentation, and incremental strategies.

Enables comprehensive evaluation of LLM applications through automated metrics, human feedback, and benchmarking frameworks.

Expert guidance on database architecture, technology selection, schema modeling, and performance optimization for scalable data systems.

Builds end-to-end ML pipelines with experiment tracking, model registry, and automated deployment using MLOps tools like MLflow and Kubeflow.

Optimizes LLM prompts using constitutional AI, chain-of-thought reasoning, and model-specific techniques to enhance application performance.

Enables building financial models, backtesting trading strategies, and implementing risk metrics and portfolio optimization using statistical methods.