MCP Circuit Designer

Technical Description

• Built an AI-powered Model Context Protocol (MCP) server enabling LLMs to design, simulate, and validate electronic circuits directly from natural language input, automating circuit design workflows for EDA applications.
• Engineered specialized tools to extend LLM functionality with domain-specific reasoning: get_component_datasheet() to fetch and convert datasheets into Markdown for precise component specifications, and get_research_paper() to retrieve academic papers using PDF-to-text pipelines.
• Integrated ngspice simulation interface to auto-generate, run, and analyze SPICE netlists, enabling evaluation of frequency response, voltage levels, and current flow for circuit validation and optimization.
• Designed intelligent design loop workflow where the LLM interprets circuit requirements, selects suitable components, generates and simulates SPICE netlists, and iteratively optimizes performance for final documentation.
• Achieved 49% improvement in circuit design quality compared to baseline LLMs, measured across component accuracy, power efficiency, signal integrity, and reproducibility metrics.
• Delivered modular, extensible architecture supporting reuse across analog, RF, and digital domains, enabling scalable circuit design workflows for Electronic Design Automation applications.

Tech Stack