Rust coding-agent proof of concept
User-selected provider/model routing, compatible custom endpoints, sandboxed repository tools, MCP, SQLite-backed sessions, parallel workers, and HTTP/SSE interfaces.
Public engineering evidence
Developer tools and retrieval systems you can inspect.
NextEleven Harness demonstrates Rust systems and agent-tooling work. PersonalRAGVault demonstrates Python retrieval, APIs, evaluation, and user-facing documentation. Both are bounded accurately: one is a proprietary public proof of concept; the other is MIT licensed.
01Implemented capabilities
Two projects, two different evidence surfaces
Public repositories and release documentation provide inspectable evidence. Repository state can change; current versions and checks belong at the linked source.
Cross-platform verification
Automated checks and cross-platform continuous integration cover documented test, lint, build, and release paths. Current results are published with the repository.
Operational documentation
Install guidance, security model, limitations, release status, provider behavior, and troubleshooting notes.
Hybrid retrieval
Multi-source ingestion, FTS5 and vector search, reciprocal-rank fusion, reranking, citations, and evaluation.
Python service surface
CLI workflows, FastAPI/SSE access, optional local models, automated tests, and user documentation.
Debuggable systems
Structured errors, reproducible behavior, documented setup, testable boundaries, and interfaces intended to help another person understand what happened.
Discuss the system or the engineering behind it.
For a technical role, implementation engagement, or focused engineering discussion, contact Sean directly and identify the project that brought you here.