3. Data, Science & AI

Generates videos from text or images using Gemini's Veo 3.1 AI model via Python SDK, supporting advanced parameters and reference images.

Strategically suggests context compaction points to maintain LLM efficiency during complex, multi-phase AI workflows.

Extracts and distributes legacy system prompt instructions to correct architectural layers (Wizard, Tool Config, Internal System) for LLM applications.

Enables persistent memory storage for AI agents to retain user preferences, project context, and learned patterns across sessions using MCP Memory Server.

Analyzes L2 market depth and trade data to detect institutional order flow patterns including absorption, exhaustion, imbalance, and sweep.

Batch-generates images via OpenAI API with random prompts and creates an HTML gallery for display.

Orchestrates AI agent workflows with SSE streaming to generate business assets in real-time.

Designs multi-agent systems for complex problem-solving by decomposing tasks and specializing agents to improve performance and scalability.

Delivers comprehensive documentation for Hugging Face Transformers library, enabling implementation of state-of-the-art NLP and LLM models.

Manages PostgreSQL database schemas with auto-healing capabilities and integrates with hybrid RAG architectures for enhanced data retrieval systems.

Filters and truncates Exa web search results to optimize token usage for AI models, enhancing efficiency in LLM applications.

Transcribes YouTube video content into text, enabling quick access to spoken information without watching the video.

Systematically reviews AI agent outputs for quality, accuracy, and completeness, identifying bugs and suggesting improvements before changes are committed.

Generates progress reports and analyzes project metrics including costs and velocity for task-graph-mcp systems.

Configures Segment CDP for customer data tracking, identity resolution, and data governance with best practices for analytics and destination integrations.

Reviews and optimizes Builder.io Fusion rule files to improve AI system effectiveness and ensure correct rule interpretation.

Converts unstructured text data (news, reviews) into structured JSON, tags, and scores via AI-powered extraction pipeline.

Imports tasks from GitHub Issues, Linear, Jira, and markdown TODOs into task-graph-mcp.

Creates, scaffolds, and validates agent skills according to official specifications, providing structure guidance and format validation for AI agent development.

Provides tutorials and documentation for PyTorch, covering deep learning model development, training, and deployment workflows.

Guides creation of Model Context Protocol (MCP) servers to integrate external APIs with LLMs, supporting Python (FastMCP) and Node/TypeScript (MCP SDK).

Provides fast, in-process analytical database operations for efficient data processing and querying within applications.

Optimizes database queries to deliver high-performance reporting with minimal latency.

Enables neuro-symbolic reasoning by integrating LLMs with symbolic solvers (Z3, Prover9) for logical problem solving and NL-to-logic translation.