Nahid H.

I build local LLM apps, AI agent tools, and integrations that connect models to your data and software. Need a private AI workspace, a document assistant, or an agent that can use your existing tools? I can help. Some of my own projects: • SmolBrain — a local knowledge workspace built in Rust, with Ollama integration, file watching, classification, and structured AI outputs. • Picoo — reusable tools for AI agents, with MCP support and workflows that handle retries, timeouts, and resuming interrupted runs. • Runspace — an agent collaboration workspace with a FastAPI gateway and local connectors for dispatching tasks and tracking runs. What I can help you build: • Local and self-hosted LLM applications • LLM API integrations and structured document extraction • RAG and document assistants with source references • MCP servers, agent tools, and workflow automation • Python/FastAPI backends and React/Next.js interfaces I also explore memory management for Mixture-of-Experts models, including routing traces and cache-policy simulations. This is research work, separate from the applications above. My backend work includes FastAPI development, authentication APIs, Stripe integrations, and Docker deployments. I can handle the application around the model too. We start with one clear use case, build a small working version, and test it on your real examples. For local models, we check hardware requirements and speed; for hosted models, response quality and running costs. Tell me what you want the AI to do, what data it should use, and where it needs to run. I'll help you work out a practical first step.