MCP Demo — Full-Stack AI Development with the Blankstate MCP Server

Objective: The primary objective of this agreement is to provide the contractor with practical, hands-on experience building a full-stack AI application in which one or more agents gather, validate, and present external data with every tool call routed through the Blankstate (BKS) MCP Server for protocol- based sensing and validation. Working alongside senior team members, the contractor will gain exposure to a modern Python and TypeScript stack (FastAPI, LangChain/LangGraph, Next.js), to sound software-engineering practice (SOLID principles and the Strategy/Adapter/Factory patterns that make data sources plug-and-play), and to the operational role of the MCP Server in observing and scoring agent behaviour. The engagement is structured to build the contractors confidence across the application's data pipeline while reinforcing professional habits of responsibility, collaboration, and clear technical communication. 2. Statement of Work Based on the contractor's background and interest in programming (Python, API consumption, data wrangling) and office software (Word, Excel), coupled with their enthusiasm for programming and demonstrated characteristics like responsibility and teamwork, the scope will include: ▪ Core Development Tasks: o Assist in building data-source adapters that wrap third-party APIs (e.g. Tavily, Firecrawl, news APIs) and expose them through the project's common DataSource interface, translating external payloads into internal Pydantic/dataclass models. o Help register new adapters with the source factory so they are plug-and-play, requiring no changes to existing agent, API, or storage code. o Support data-wrangling work: cleaning, normalizing, and validating data retrieved from external sources before it is stored. o Assist with consuming and parsing external APIs as part of adapter development. ▪ Supported Learning & Contribution Tasks (Under Direct Guidance): o Contribute, with close supervision, to specific FastAPI endpoints (e.g. /api/v1/data, /api/v1/sources) and their Pydantic request/response models. o Assist senior engineers with portions of the LangChain/LangGraph agent flow and the integration points with the BKS MCP Server, observing how tool calls are sensed and validated before and after execution. o Help write PyTest unit tests for adapters, strategies, and the factory, using mocks at external boundaries so tests remain deterministic and offline. ▪ Administrative & Research Support: o Participate in relevant team meetings and stand-ups. o Help maintain project documentation (e.g. setup notes, adapter-authoring guides) using Word, Excel, or the designated documentation platform. o Conduct research on specific technical concepts (e.g. MCP, SOLID, the design patterns in use, data-privacy/GDPR considerations) as required. 3. Deliverables Expected deliverables during the agreement period include (but are not limited to): ▪ One or more completed data-source adapters implementing the common DataSource interface, with corresponding factory registration. ▪ Accompanying PyTest unit tests for assigned adapters, strategies, and factory logic. ▪ Guided contributions to specific FastAPI endpoints and their Pydantic models. ▪ Cleaned and normalized sample datasets demonstrating the validation pipeline. ▪ Updated project documentation (setup, adapter-authoring, or related sections). ▪ Regular progress updates to the reporting manager. 4. Candidate Suitability & Support We recognize the contractor’s foundational knowledge in programming and their aptitude for learning. While direct experience in documenting complex AI systems, architecture, or databases may be limited, this agreement is designed to provide growth opportunities. Blankstate is committed to providing mentorship, clear guidance, and necessary resources to support the contractors in developing the skills required for the more technical aspects of the role. Their proficiency in Word and Excel will be immediately applicable to many tasks.