AI Systems Engineering
End-to-end model engineering across training, evaluation, quantization, inference serving, and application integration.
Back to the team
Technology FocusLLM Systems & AI Infra
Current RoleFounder & CEO
AI Engineering · System Architecture · Leadership
Zhewen Li李 哲闻
Founder & CEO · AI Systems Architect
I connect enterprise software engineering with model training, inference optimization, and AI product delivery. My current focus includes production-grade vLLM and SGLang serving, RAG, agentic systems, and multimodal applications.
Tokyo · ShanghaiChinese · Japanese · EnglishWenku Technology
Professional Profile
Turning frontier model capabilities into dependable business systems
I am a hands-on technology leader with experience across system architecture, product delivery, distributed platforms, data systems, and generative AI. Starting from business outcomes, I design end-to-end solutions spanning data governance, training and evaluation, inference optimization, application orchestration, and cloud-native operations.
I stay close to critical prototypes and architecture decisions—balancing model quality with latency, throughput, cost, observability, and long-term maintainability.
Enterprise Architecture
Scalable distributed, data, and cloud-native systems designed for availability, concurrency, and complex collaboration.
Product & Delivery
Connecting business, design, and engineering to turn technical prototypes into measurable enterprise products.
AI Technology Stack
A complete capability stack from training to online inference
Beyond calling model APIs, I build evolvable AI systems around data, models, compute, serving, and applications.
Training & Alignment
Efficient domain adaptation, alignment, and evaluation with enterprise data
Inference & Serving
System optimization for throughput, latency, memory efficiency, and reliability
AI Applications & Agents
Orchestrating model capabilities into traceable and testable workflows
Cloud-Native & Full Stack
Backend, frontend, and infrastructure capabilities for continuous AI delivery
Inference Engineering
High-performance LLM inference
Selecting and tuning inference engines around model size, context patterns, concurrency, and cost objectives.
vvLLM
OpenAI-compatible serving, PagedAttention, continuous batching, quantized deployment, and multi-GPU scaling for high-throughput inference.
SSGLang
RadixAttention, prefix caching, structured generation, and efficient execution for complex agentic and multimodal workflows.
Career Architecture
A dual-track career
A long-term entrepreneurial and technical leadership track, complemented by hands-on enterprise roles and industry projects.
Why do some dates overlap?
Wenku Technology is the continuous entrepreneurial track. Other entries represent concurrent enterprise roles, advisory work, or project engagements. The two-track format makes those responsibilities and timelines explicit.
01
Entrepreneurship & leadershipContinuous operation · Long-term ownership
Apr 2020 — Present
Wenku Technology
Founder & Chief Executive OfficerLeading technology strategy, product direction, and core architecture for AI-native enterprise solutions. Deeply embedded with the RTE (Release Train Engineer) team, partnering with product and delivery teams to introduce AI Agents across requirements analysis, architecture, coding, testing, and operations, and to establish an AI Native development model. The technology portfolio spans LLM fine-tuning and high-performance inference, RAG, MCP, agentic long-term memory and evaluation, computer vision, OCR, and cloud-native distributed systems. Driving platformization, engineering standards, observability, and governance to turn project knowledge into reusable organizational intelligence, while remaining hands-on in critical architecture, prototyping, and production implementation.
02
Concurrent professional practiceEnterprise roles · Projects and industry depth
From Aug 2022 · Concurrent
NTT DATA
Deputy Manager / System ArchitectContributed to centralized payment platforms and enterprise architecture, focusing on transaction security, availability, scalability, and cross-team delivery.
From Apr 2021 · Concurrent
Sense AI
Data ScientistBuilt data-platform and analytics capabilities that turned fragmented information into reusable data assets and decision insight.
From Nov 2020 · Concurrent
Accenture
Senior AnalystSupported enterprise data and digital initiatives through business analysis, data modeling, and system delivery for large-scale transformation.
Education
Academic background
Arizona State University
Computing and TechnologyAdvanced study in scalable computing, enterprise cloud applications, big data, machine learning, and modern system architecture.
Tokyo Institute of Technology
Natural Language ProcessingFocused on language understanding, text analytics, machine learning models, and applied NLP engineering.
Soochow University
Computer ScienceBuilt a rigorous foundation in computer science, operating systems, networking, and computer architecture.
Credentials
Professional certifications
Jun 2025
Big Data Professional Certification
Dec 2024
AI and Machine Learning Professional Certification
Oct 2024
AWS Certified Solutions Architect – Professional
Dec 2020
Oracle Certified Java Programmer Gold SE 8
Aug 2020
Google Data Analytics Certification
Feb 2016
Japanese Language Proficiency Test N1
Build What Matters
Start a conversationBring AI into real business systems
If you are planning an enterprise AI program, inference platform, or complex system modernization, let’s discuss the technology path and delivery strategy.