Resume

Socials:

Location:

China

Position:

Datacenter technician, DevOps, Data analysis, Data engineering, Cybersecurity, Research assistant, DevOps, DevSecOps

Summary:

Network and system administration, on-prem, cybersecurity, data engineering, machine learning, DevOps, data analytics. Focused on open-source solutions. I like the concept of a data-driven world and getting insights from data, enjoy building complex systems. Love learning new things and adding value to other people’s lives.

Projects:

On-premise GPU cluster | 2022.9 - Present

Designed and built on-premise multi-node multi-GPU hyper-converged cluster, including surrounding infrastructure, cybersecurity and helpdesk. 10GbE fiber optics in LACP bonds and 200G Nvidia InfiniBand. Proxmox VE, CEPH with RBD erasure coding pools, FreeIPA central authentication, LXC containers, Docker.

Project outcomes: A system with reliable redundant high-speed storage for a large number of small files (audio and text) that allows running data analysis, data processing and training deep learning models. HCI lowers the total cost of ownership, and increases flexibility and scalability. Increased flexibility of resource allocation allows for more efficient hardware utilization.

GPU docker (DevOps)

Deployed a very flexible, straightforward (no hacks) and scalable system for non-exclusive allocation of GPU compute and VRAM resources between multiple docker installations within one cluster node while maintaining a certain degree of isolation using unprivileged LXCs. Project outcomes: multiple GPUs are shared among multiple isolated docker instances and can be assigned to docker containers individually. GPU compute overcommitment leads to more efficient utilization and less time spent on idle. Writeup 1, writeup 2

Horizontal scaling

Rewriting single process CPU and GPU tasks Python code to use multiprocessing and multithreading to optimize compute load and shorten task execution time. Tasks range from text and audio signal data engineering to deep learning models inference.

My other Projects

Papers

Shuai Wang, Yuzhu Liang, Tianheng Wang, Wei Huang, Ke Xu, Aleksandr Mitkov, Shuting Yuan, Jun Ding, Lining Wang, Yongdao Zhou, Quansheng Xia, Qibin Ran, Unraveling the influence of essential climatic factors on the number of tones through an extensive database of languages in China, Journal of Language Evolution, Volume 9, Issue 1-2, January/July 2024, Pages 29–48, https://doi.org/10.1093/jole/lzae009
陆旭,Aleksandr Mitkov,冉启斌.从词向量计算看汉语名词和动词的关系[J].语言教学与研究,2024,(03):57-67.《语言文字学》2024年第10期全文转载。《词汇学理论与应用（十二）》
Masters thesis: Using machine learning in an experimental study on the Chinese tones acquisition by Russian students. CNKI:CDMD:2.1022.487181

Skills (most recent experience):

Computer system and network administration:

Setting up and maintaining multi-purpose Linux servers (Ubuntu, Debian)
Setting up and maintaining GPU/compute/storage clusters (Proxmox VE, CEPH)
GPU-enabled virtualization (LXC, QEMU/KVM)
DevOps (docker, SQL, NGINX), on-prem LLM (n8n, ollama, hf)
Data management, backup, disaster recovery
Cybersecurity, Pentesting
wireless and wired networking (including NVIDIA InfiniBand), former MTCNA
Helpdesk (Windows/MacOS/Linux, software, data management, programming), Documentation
Computer hardware troubleshooting and repair
FPV drone building (from scratch) and piloting, Betaflight/ELRS/EdgeTX programming
Raspberry Pi projects

Data engineering and deep learning:

Signal processing
Writing code and training deep learning models (Tensorflow, Pytorch)
Model performance metrics engineering and analysis, experiment tracking (mlflow)
English/Chinese natural language processing, embeddings

Programming:

Python (NLP, DL/ML and scientific frameworks, visualization), Shell script, Jupyter, git
Entry-level full-stack (JavaScript, React, Python, FastAPI) development

Training and education:

Second language education

Language:

Chinese: advanced
English: advanced
Russian: first language

Certifications:

C1D driving license, 轻型民用无人驾驶航空器安全操控理论培训合格证明, HSK5, HSKK, NVIDIA Fundamentals of Accelerated Computing with CUDA Python, MTCNA (expired)

Interests: autonomous driving. TryHackMe, HackTheBox.

Education:

Year	Degree	University	Major
2022-	PhD	Nankai University	Digital Humanities
2020-2022	Master of Arts	Hebei Normal University	International Education, Teaching Chinese to Speakers of Other Languages
2016-2020	Bachelor of Arts	Hebei Normal University	International Education, Teaching Chinese to Speakers of Other Languages

Working Experience:

Year	Position	Responsibilities
2003-2016	Network and System Administrator	System and Network administrator. Responsibilities include maintaining on-premises servers running with terminal servers inside Microsoft Hyper-V instances, wired and wireless networking (UBNT, Mikrotik), Linux, Windows/Mac systems, thin clients, pfsense, ISA server.