EP.01 · THE ENGINEER BEHIND THE CODE

DEVELOPMENT LOG / CHINA → GLOBAL

ZICONG “CASPIAN” ZHANGSoftware Development Engineerソフトウェア開発エンジニア

I build production-grade backend systems and applied AI infrastructure—from high-concurrency services and event-driven pipelines to retrieval-grounded LLM applications.

Explore projects View experience

PythonJavaGoKafkaAI / LLM

VISUAL ID / 凪既知を越えて、未知へ。

Beyond the known. Into the unknown.

ROLESOFTWARE ENGINEER

ORIGINHANGZHOU, ZHEJIANG

CURRENT LOCATIONHOUSTON, TEXAS

ONSITE · REMOTE · RELOCATE

ROLESOFTWARE ENGINEERBACKEND · GENERAL SDE · AGENT

ORIGINHANGZHOU, ZHEJIANG

CURRENT LOCATIONHOUSTON, TEXAS

WORK LOCATIONONSITE · REMOTE · RELOCATE

VERIFIED ENGINEERING PROFILE

ZICONG ZHANG

Software Engineer / AI Engineer

PROFILE IDZZC / NAGI-01DISCIPLINEBackend · General SDE · AgentsWORK MODEOnsite · Remote · Relocation

I am a Master of Computer Science candidate at Rice University with a background in distributed systems, cloud infrastructure, and production LLM applications. I enjoy turning complex requirements into reliable, measurable software.

EDUCATION RECORD

Aug 2025 – Dec 2026Rice University

Master of Computer Science

GPA: 3.9/4.0

Sep 2021 – Jun 2025Jiangsu University

Bachelor of Engineering in Computer Science and Technology

GPA: 3.67/4.0 · Top 10%

LOCATION: Houston / Hangzhou
FOCUS: Software · AI · Agents
STATUS: Open to opportunities

MISSION DATA DECRYPTED

May 2026 – Aug 2026 · Hangzhou, Zhejiang, China

FABU Technology (FABU.ai)

Agent Development Intern (Software Development Intern)

OBJECTIVE

Built construction-safety AI systems combining multimodal retrieval, vision LLMs, document ingestion, and grounded regulation Q&A.

IMPACT

Embedded a 5,000+ labeled violation-image library in Milvus with Qwen3-VL embeddings and retrieved the top five similar cases as grounding context.
Served the detection flow through FastAPI and Docker with sub-200ms retrieval, p99 under 3 seconds, and 92% detection accuracy.
Ingested 1,000+ compliance PDFs and built a LangGraph regulation agent reaching 89% accuracy against a human-expert golden evaluation set.

FastAPIMilvusLangGraphQwen3-VLRAG

03 / PROJECTS

PROJECT ARCHIVE / HOLOGRAPHIC DATA RAIL

0102SCROLL OR USE ARROWS · SELECT TO READ

ENGINEERINGCORE06 MODULES ONLINE

ACTIVE MODULE / 02

AI / ML

PyTorchTensorFlowQLoRALangChainLangGraphRAGLLM Evaluation

USED INChatbot Arena · multimodal retrieval · regulation agent

‹‹‹ HOVER TO UNLOCK · CLICK TO BREAK ›››

EP.06 / Contact

Let’s build something meaningful.

Based in Houston, Texas. Targeting Backend SDE, General SDE, and AI agent engineering opportunities in the United States. Open to onsite and remote roles, including relocation.