Skip to content
View MavCui's full-sized avatar
  • APAC

Block or report MavCui

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
MavCui/README.md

Maverick Cui — AI Solution Architect at ByteDance

AI Solution Architect at ByteDance. Founder before that.

For four years I've worked on the unglamorous half of AI — agent runtimes, device memory, risk decisioning — at Baidu, Xiaomi and SAS. My own ventures have gone through Antler and other top-tier accelerators.

The conclusion hasn't changed: the model is never the hard part. Everything around it is. That's what I build now for growth-stage AI companies — LLM, video and world-model workloads, from demo to deployed.


Experience

Eight roles from 2023 to now: ByteDance, Baidu, Xiaomi, TCL, SAS, and earlier work at China Merchants Group, China Mobile, Deloitte and BMW

Plain text version

ByteDance · Volcano Engine — AI Solution Architect · 2026.09 – now Solution architecture for growth-stage AI companies, connecting LLM, video and world-model workloads with infrastructure and delivery requirements.

Baidu · Overseas Innovation — AI/ML Systems Architect, Medo.dev · 2026.04 – 2026.09 Founding member of the overseas team; four regions in three months. Multi-model routing, tool orchestration, multi-agent workflows, end-to-end solution delivery.

Xiaomi · Technology Committee — AI Engineer, Product Architect · 2025.11 – 2026.03 GUI agents, super-Xiaoai memory and proactive intelligence, across the first AIOS in China to schedule phone, AIoT, car and PC as one surface.

TCL · Honghu AI Lab — Applied AI Engineer, AiMe Robot OS · 2025.07 – 2025.11 Personality-layer evaluation and interaction strategy for a companion robot: a nine-dimension experience framework plus automated LLM personality evaluation.

SAS Institute · Sydney — AI Product Owner, AMRSS · 2025.02 – 2025.07 Australia's first AI credit-risk platform — feature engineering, model inference, decision policy and audit trails in one configurable chain. Deployed at five banks.

China Merchants Group — Data Scientist · 2024

China Mobile R&D — AIGC imaging platform, Stable Diffusion and LoRA training · 2024

Deloitte Hong Kong — Technology consulting · 2024

BMW China — IT PMO, next-generation iX3 and 5 Series cockpit · 2023 – 2024

Recognition

NVIDIA Hackathon · three championships across two years · project owner
风御 — edge AI network security on BlueField DPUs, closing the loop from detection to autonomous defense. And an AI-native org platform for overseas enterprises: control plane, persistent workspace, RSI self-evolution.

China International Innovation Competition 2025 · national gold · project owner
Neurowear (patent) — on-device multimodal brain-health management, taking it from a single hospital check to continuous daily observation. DenseFormer-MoE on wearable signals, with long-horizon trend tracking and expert review.

China International Innovation Competition 2026 · Oceania champion · project owner
Synkora (patent) — the first compute–energy co-optimization model for AI factories. When to run a job, which site, which GPU, against power prices and delivery deadlines.

AI & Engineering Stack

Engineering AreaTechnologies & Ecosystem
Agent Architecture & Infrastructure LangGraph   LangChain   LlamaIndex   Model Context Protocol
LangGraph · LangChain · LlamaIndex · MCP · A2A
LLM Engineering & Serving PyTorch   Hugging Face   vLLM   NVIDIA
PyTorch · Hugging Face · vLLM · SGLang · LiteLLM · TensorRT-LLM
Retrieval, Memory & Evaluation PostgreSQL   Redis   Elasticsearch   Langfuse
PostgreSQL · pgvector · Redis · Qdrant · Langfuse · Ragas
Cloud-Native AI Infrastructure AWS   Azure   Google Cloud   Docker   Kubernetes   Volcano Engine
AWS · Azure · Google Cloud · Volcano Engine · Docker · Kubernetes
Full-Stack Engineering Python   TypeScript   Go   React   Next.js   FastAPI
Python · TypeScript · Go · React · Next.js · FastAPI
More engineering tools
Engineering AreaTechnologies & Ecosystem
Post-Training & Agent RL Hugging Face Transformers, PEFT and TRL   PyTorch
Transformers · PEFT · TRL · verl · Ray
iOS & On-Device Intelligence Swift   Xcode   Apple platform development
Swift · SwiftUI · Xcode · Core ML · Foundation Models · MLX
Data & Developer Tools MySQL   Git   GitHub Actions   Linux   Figma
MySQL · Git · GitHub Actions · Linux · Figma
Models & APIs OpenAI   Claude   Gemini   Doubao   DeepSeek   Qwen
OpenAI · Claude · Gemini · Doubao · DeepSeek · Qwen

Certifications

42+ CERTIFICATIONS · 10 CREDENTIALS SHOWN

Google Cloud Professional Cloud Architect    Google Cloud Generative AI Leader    AWS Solutions Architect – Professional    AWS Generative AI Developer – Professional    Azure Solutions Architect Expert

Agentic AI Business Solutions Architect    Oracle AI Vector Search Professional    Oracle OCI Generative AI Professional    Oracle MySQL 8.0 Database Administrator Professional    NVIDIA Certified Professional: Gen AI LLMs

View credential details

Education & Community

UNSW Sydney · Bachelor of Computer Science · 2023 – 2026 · Excellence Scholarship
President, Google Developer Group UNSW · Founder & Executive President, UCS


LinkedIn GitHub

WeChat · -MrCui- — bring a question, or just come say hi.

Pinned Loading

  1. EnterOS-AI/enter-os-core EnterOS-AI/enter-os-core Public

    The first operating system for AI agents — run heterogeneous agents as one governed, memory-driven, production-ready organization.

    Go 47 3