Open to Work — AI & Engineering Roles | Original
Note: This post is intended for friends and connections outside my current workplace. If you are a colleague or manager from my current employer, please kindly ignore this post.
Hi, I’m Zhiwei.
I’m a software engineer with 12 years of experience spanning iOS, Android, frontend, backend, and AI. I’ve worked across live-streaming startups, cloud computing platforms, China internet products, and financial systems at Singapore and Hong Kong banks. I’m actively moving into AI engineering — training GPT-2 with nanoGPT on H200 GPUs, consuming ~1.5 billion tokens via LLM APIs (~500M last month alone), building personal agents and automation tools. I ranked top 6% in AI assistant usage globally at my current employer. I want a role where AI, agents, and LLM work are the main job — not a side project.
📌 Highlights
- 12 years of engineering experience across iOS, Android, frontend, backend, and AI
- Worked across diverse domains: live-streaming startup, cloud computing startup, China internet products, and financial systems / digital banking apps of Singapore & Hong Kong
- Born 1995, Chinese citizen, based in Guangzhou; IELTS 6.5 (Reading 8.5)
- Enrolled at Beijing Forestry University (2013–2014), dropped out; pursuing Associate Degree via self-study — 13 of 17 courses passed
- 2011 NOIP Guangdong province, top 300 provincially; 1st prize (Guangzhou round), advanced to provincial round
- Founded Fun Live (live-streaming app, 30,000 users) and Square Root Technology; reached 3M CNY revenue with 600K CNY profit in 2018
- Received a job offer of ~384,000 CNY/year (~32K/month) around 2023, without a bachelor’s degree
- 10+ open-source GitHub projects (500+ commits each); ~400 technical blog posts; ~8,000 AI answer notes
- Trained GPT-2 124M with nanoGPT twice (RunPod H200, DigitalOcean H100, home RTX 4070); ~1.5 billion tokens consumed via OpenRouter & other LLM providers in the past year (~500M last month)
- Built ww (CLI toolkit), iclaw (AI coding REPL for enterprise environments), and zz (dataset processing & training utilities for ML projects)
- Self-taught researcher — 3 papers on natural vision restoration; reversed myopia 350 → 250 diopters over 3 years
🤖 AI Projects
Model Training
- Trained GPT-2 124M from scratch twice using nanoGPT on the FineWeb dataset — once on a RunPod H200, once on a DigitalOcean H100; also ran experiments on a home server with an RTX 4070.
- Used zz (dataset processing & training utilities) to download, extract, and analyze FineWeb data; scripts cover dataset download, parquet extraction, training duration calculation, and metric evaluation.
- Completed Machine Learning Specialization (DeepLearning.AI & Stanford University) and Deep Learning Specialization (DeepLearning.AI) on Coursera.
Personal AI Projects
- jekyll-ai-blog — Built an AI-powered blog platform with automated multi-language translation, Google Cloud TTS audio generation, XeLaTeX PDF/EPUB pipelines, and GitHub Actions workflows.
- lzwjava.github.io — Personal blog and knowledge base with ~400 original posts and ~8,000 AI answer notes; ~70,000 page views in the past month (Cloudflare Analytics), with Singapore as the top visitor country.
- ww — Cross-platform CLI toolkit for developer productivity: git workflows with AI commit messages (Gemini Flash), image/PDF processing, web search, GitHub Copilot chat, system utilities, and LLM-powered helpers.
- iclaw — Terminal AI agent (REPL) that codes, searches, and runs shell commands autonomously. Supports GitHub Copilot (OAuth) and OpenRouter; designed for both personal machines and locked-down enterprise environments with no IDE plugins required.
- zz — Dataset processing and training utilities for ML projects: FineWeb dataset download/extraction, training log analysis, and evaluation scripts used during GPT-2 training runs.
- live-server — Used OpenClaw (AI coding agent) to modernize a previous startup project: Dockerized the application, upgraded CodeIgniter and Vue to current versions, and modernized the full stack.
Open Source Contributions to Others’ Projects
- Tree_Of_Thought (1 PR) — Contributed to a friend’s Tree-of-Thought reasoning system; added OpenAI-compatible requester and
python-dotenvconfig.
LLM API Usage
- Consumed ~1.5 billion tokens via OpenRouter and other LLM providers in the past year; ~500 million tokens in the past month alone — reflecting deep, daily hands-on use of LLMs for coding, research, and automation.
🔍 Why I’m Looking
My current job is going well — I’ve been with the same vendor-bank engagement for over a year, and my contract between my vendor and the bank has recently been extended for another year. I’m performing strongly: top 20% among contractors at my vendor, and top 6% in AI assistant usage globally at my employer.
My primary motivation for looking is becoming more AI-centric. I have technical idols — Yin Wang, Andrej Karpathy, Wenfeng Liang, Greg Brockman — and I want to grow in the direction they represent: deeply technical, AI-first, and building things that genuinely help companies and users. My current role involves AI tooling at the edges, but I want a position where AI, agents, and LLM systems are the core of the work, not a side activity. That said, given my reality, contractor or permanent positions at big banks that combine AI with backend or full-stack engineering are welcome too — I know that space well and can contribute immediately.
Compensation improvement is a secondary but real motivation. My current package is below where I’d like to be, and I want a role that better reflects my experience and output.
Beyond both of these, regularly engaging with the market is simply good practice — it keeps perspective sharp and ensures I’m aware of what’s available. The bank has a contractor-to-permanent conversion policy, but slots are limited and another senior peer is ahead in the queue. Looking externally is the more realistic path forward.
Full-stack and backend roles are welcome. My strong preference is for positions where AI, agents, or LLM work are central — I’ve been deeply invested in this space outside of work hours and want a role that matches that direction.
🏢 Work Experience
| Company | Role | Period |
|---|---|---|
| TEKsystems → HSBC Bank (contractor) | AI Engineer | 2025.02 – Present |
| Freelancer | ML & AI Projects | 2023.08 – 2025.01 |
| Farben Information (outsourced to HSBC PayMe) | Backend Engineer | 2022.11 – 2023.07 |
| Beyondsoft (outsourced to DBS Bank) | Backend Engineer | 2021.12 – 2022.11 |
| Freelancer | Full Stack & Consulting | 2020.01 – 2021.11 |
| Square Root Technology (Fun Live app) | Founder & Full Stack Engineer | 2016.07 – 2019.12 |
| CodeReview.cn | Co-founder & Full Stack Engineer | 2015.11 – 2016.07 |
| LeanCloud | Software Engineer | 2014.07 – 2015.11 |
TEKsystems → HSBC Bank, AI Engineer (Contractor), 2025.02 – Present
- AI Engineer outsourced to HSBC Bank, leveraging Copilot and its API to accelerate backend development for the Finance Transformation Platform in HSBC’s Enterprise Technology Department.
- Maintained financial data processing features (import, validation, export) and enhanced submission/approval workflows; gained hands-on experience with accounting, ledger, and banking systems.
- Participated in the full development lifecycle — local development through UAT to production. Helped decommission legacy WebSphere apps, automated releases with Ansible and Jenkins, and assisted in a major Angular upgrade.
- Led integration and API testing, using the Copilot API to auto-generate ~70 test cases covering Spring Filters, Python unittest, JSON truncation, prompt engineering, and regional endpoints.
- Built a personal AI agent layer — 20 customized agents, 400 reusable scripts, and 1,100 Copilot-written guides — to automate scripting, logging, and documentation; ranked top 6% in Copilot usage globally (premium requests metric).
- Joined HSBC’s internal AI community; earned a Contribution Award for the AIPlayer project. Explored AI after hours: nanoGPT training on H200/RTX 4070, personal projects via OpenRouter and Claude Code, and study of llama.cpp, Transformers, and reasoning techniques.
- Stack: Java, Spring, IBM Db2, Maven, Angular, Python, HashiCorp Vault, Ansible, Control-M, IBM WebSphere Liberty Profile, Copilot.
🎓 Education
| School | Major | Period |
|---|---|---|
| Guangdong University of Foreign Studies | Computer Application, Associate Degree (Self-study, 13/17 courses passed) | 2022.10 – Present |
| Beijing Forestry University | Digital Media Arts, Bachelor (Dropout) | 2013.09 – 2014.06 |
| Guangzhou Yuyan Middle School | Science Track | 2007.07 – 2013.06 |
💼 Roles I am open to
I am flexible and open to multiple engineering directions:
- AI Engineer (Applied LLM / NLP / AI systems)
- Machine Learning Engineer (Python / LLM / data-driven systems)
- Backend Engineer (AI-related systems or platform engineering)
- Full-Stack Engineer (AI product integration / SaaS platforms)
- Distributed Systems / Platform Engineer (AI infrastructure)
- Full-Stack Java Engineer (Global Banks / Financial Systems)
- Senior Software Engineer (IC track, any of the above domains)
- Staff Engineer (technical leadership, cross-team scope)
- Tech Lead (half-lead, half-hands-on; previously managed 4 FTEs and ~20 part-time contractors at my startup)
- Member of Technical Staff (MTS) — individual contributor at any level
- Early Founding Member / Founding Engineer at an AI-focused startup
I am not limited to pure research roles — I focus on building real systems and shipping products.
🏢 Work Arrangement
Open to:
- Full-time permanent roles
- Contract
- Startup environments (fast iteration, product ownership)
- Large enterprises / big tech companies
🌍 Location Flexibility
- Shenzhen / Guangzhou — preferred; open to 5 days/week onsite
- Other cities in China — prefer 2–3 days/week in office; open to 5 days/week if salary is more competitive
- Hong Kong — open, but would require work visa sponsorship
- Singapore — open, but would require work visa sponsorship
- Remote — open, global teams welcome
💰 Compensation Expectation
Flexible depending on scope, responsibility, and growth potential.
- 🇨🇳 Mainland China: 30,000 – 45,000 CNY/month
- 🇭🇰 Hong Kong: ~40,000 HKD/month or above
- 🇸🇬 Singapore: ~9,000 SGD/month or above
🧠 Technical Background
- Strong backend engineering: Java, Spring Boot, distributed systems, microservices
- Full-stack development: Vue / React / Angular / frontend + backend integration
- Cloud platforms: AWS / Azure / Alibaba Cloud
- Databases & infrastructure: MySQL, Redis, logging, monitoring systems
- API systems and large-scale service design
- Hands-on LLM experimentation (nanoGPT, small model training, prompt + API systems)
- AI application development using LLM APIs
🌐 Language Ability
- Mandarin Chinese (native)
- English (IELTS 6.5, professional working proficiency)
- Cantonese (conversational)
- Comfortable working in both English and Chinese environments
🏅 Certificates
- IELTS Academics, Score 6.5
- Machine Learning Specialization — DeepLearning.AI & Stanford University
- Deep Learning Specialization — DeepLearning.AI
- Developing on AWS — AWS Training
- Beijing Forestry University Dropout Certificate
- China Higher Education Official Report (Bachelor, Dropout Status)
- GitHub: https://github.com/lzwjava
- Blog: https://lzwjava.github.io
- Portfolio: https://lzwjava.github.io/portfolio-en
I’m open to discussing any roles where AI, systems engineering, and real product development intersect.