research-local-xiaomi
Local research workflow for turning web evidence into structured reports with planning, critique, usage tracking, and reproducible outputs.
Shows research pipeline design around local models and agentic review loops.
Work
A curated view of public, private, client, academic, and experimental work. The homepage stays selective; this page gives the broader system.
Capability groups
Local research workflow for turning web evidence into structured reports with planning, critique, usage tracking, and reproducible outputs.
Shows research pipeline design around local models and agentic review loops.
Full-stack AI employee prototype for agribusiness: RAG, local LLM integration, crawler, FastAPI backend, PostgreSQL, Qdrant, and Docker Compose.
Demonstrates end-to-end AI system architecture from frontend to vector search and local inference.
Local browser agent that controls visible Chromium through Playwright using ARIA snapshots, structured tool calls, memory, and logging.
Demonstrates practical agent loop design for browser automation.
Business MVP for IFRS/MSFO-style financial workflows: data import, reports, charts, and export-oriented interface.
Shows ability to turn a domain workflow into a usable internal tool.
Private client work: MCP tools for Redmine and a role-aware VK Teams RAG bot over Wiki.js and PDF regulations.
12 Redmine tools.
IDP architecture for long, unstructured documents: OCR, LLM orchestration, Smart Anchor extraction, async backend, queues, caching, and S3 storage.
Estimated form-filling time reduced from 40 minutes to about 3 minutes.
Local analysis pipeline for exported chat data: audit, ingest, SQLite storage, staged extraction, and reports.
Shows staged extraction and local reporting over exported data.
Local analytics MVP for cleaning, monitoring, and forecasting sales data.
Shows compact analytics MVP delivery for a domain workflow.