8000 xugangqiang (xugangqiang) · GitHub
[go: up one dir, main page]

Skip to content
View xugangqiang's full-sized avatar
  • HP
  • Shanghai

Block or report xugangqiang

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
xugangqiang/README.md

Hello, World

Technical Architect / Team Lead / Backend Expert

     ██╗ █████╗  ██████╗██╗  ██╗    ██╗  ██╗██╗   ██╗
     ██║██╔══██╗██╔════╝██║ ██╔╝    ╚██╗██╔╝██║   ██║
     ██║███████║██║     █████╔╝      ╚███╔╝ ██║   ██║
██   ██║██╔══██║██║     ██╔═██╗      ██╔██╗ ██║   ██║
╚█████╔╝██║  ██║╚██████╗██║  ██╗    ██╔╝ ██╗╚██████╔╝
 ╚════╝ ╚═╝  ╚═╝ ╚═════╝╚═╝  ╚═╝    ╚═╝  ╚═╝ ╚═════╝





What I Do

20+ years building high-concurrency backend systems across banking, e-commerce, recommendation systems, and AI infrastructure.

  • Currently at InfiniFlow — leading RAGFlow's Python→Go platform migration: 200+ merged PRs covering the document REST API layer, task executor, ingestion pipeline (PDF/DOCX/PPT/email parsers, chunker, tokenizer, embedding), knowledge-graph retrieval, and the DeepDoc inference stack.
  • Previously at Morgan Stanley · NVIDIA · Coupang · iQIYI, from enterprise Java delivery to architecting 0→1 platforms.

🔧 Recent Upstream Work

  • ragflow #18731: DeepDoc document analysis as a native in-process Go backend over statically-linked ONNX Runtime (cgo, dlopen-resolved), with golden-test parity against the Python reference — 3.5px coordinate tolerance, race-detector-verified concurrency, CI drift gate (merged)
  • ragflow #18688 · #18707 · #18727 · #18740 · #18789: PDF & table parsing hardening — cross-page table merge, colspan assembly, chunker position handling (merged)
  • ragflow #19247: multi-platform native builds (linux/amd64, linux/arm64, darwin) (merged)
  • office_oxide #116: table extraction from legacy binary .doc per the [MS-DOC] spec, incl. merged-cell structure reconstruction — shipped as a core feature of v0.1.9 (merged)
  • office_oxide #120: list & tab-stop extraction from legacy .doc (ilfo-keyed list detection, sprmPChgTabs decoding) (merged)

Core Competencies

Architecture — High-concurrency distributed systems · Cross-language platform migration with behavioral-parity guarantees

Data — Real-time & offline pipelines · Document parsing & ML inference serving

Leadership — Cross-team coordination (fluent English) · Performance & cost optimization


Experience

InfiniFlow · Backe 8000 nd Architect · 2026 — Present
NVIDIA · Engineer · 2024 — 2025
Coupang · Architect · 2020 — 2023
iQIYI · Recommendation Tech Lead / Architect · 2019 — 2020
Morgan Stanley · Asia Emerging Markets Tech Lead · 2014 — 2019
HP · EMC · Wipro · Senior Engineer / Team Lead · 2004 — 2014


Education

Zhejiang University · Software Engineering · Bachelor's · 2000 — 2004


📊 GitHub Stats



Get in Touch


Pinned Loading

  1. infiniflow/ragflow infiniflow/ragflow Public

    RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

    Go 93.7k 11.3k

  2. yfedoseev/office_oxide yfedoseev/office_oxide Public

    The fastest Office document library for Python, Rust, Go, JS/TS, C# and WASM. DOCX, XLSX, PPTX, DOC, XLS, PPT. Up to 100× faster than python-docx/openpyxl/python-pptx. 100% pass rate on valid Offic…

    Rust 162 25

  3. deepseek-harness deepseek-harness Public

    Forked from deepseek-ai/deepseek-harness

    DeepSeek Harness: Everything is a Plugin.

    TypeScript

  4. pdf_oxide pdf_oxide Public

    Forked from yfedoseev/pdf_oxide

    The fastest PDF library for Python and Rust. Text extraction, image extraction, markdown conversion, PDF creation & editing. 0.8ms mean, 5× faster than industry leaders, 100% pass rate on 3,830 PDF…

    Rust

0