Career Action Plan for Zhiwei Li | Generated by AI
Question: Based on Zhiwei Li’s resume (lzwjava.github.io/resume-en) and job search post (lzwjava.github.io/job-search-en), what are the recommended things to do in the next 12 months?
Answer:
Based on a thorough reading of Zhiwei Li’s resume and job search post, here is a comprehensive, prioritized action plan for the next 12 months.
🎯 Overview of Context
Zhiwei is an AI Engineer at HSBC (via TEKsystems), actively looking for a more AI-centric role. He has 12 years of experience spanning iOS, Android, frontend, backend, and AI. He’s ranked top 6% in AI assistant usage at his current employer and has consumed ~1.5 billion tokens via LLM APIs in the past year. His compensation target is 30,000–45,000 CNY/month in mainland China.
📚 1. Complete the Associate Degree (Highest Priority)
Zhiwei is pursuing an associate degree through self-taught education and has passed 13 of 17 courses. He still needs to pass 5 written exams (Computer Network Technology, Switching and Routing Technologies, Mao Zedong Thought and Theory of Socialism with Chinese Characteristics, College Chinese, An Overview of Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era) and 5 practical exams (Fundamentals of Computer Programming, Computer Fundamentals and Application Technology, Database and Its Applications, Web Design and Production, Mobile Application Development Technology).
Action: Schedule and sit for all remaining 10 exams within the next 12 months. Getting this credential removes a key CV weakness and may unlock higher salary bands.
🤖 2. Deepen AI/LLM Engineering Skills
Zhiwei wants a role where AI, agents, and LLM systems are the core of the work, not a side activity. He has technical idols — Yin Wang, Andrej Karpathy, Wenfeng Liang, Greg Brockman — and wants to grow in the direction they represent: deeply technical, AI-first, and building things that genuinely help companies and users.
Actions:
- Train larger models (GPT-2 Medium/Large or GPT-3 scale) using nanoGPT or HuggingFace Transformers, not just the 124M model already done.
- Go deeper into Reinforcement Learning from Human Feedback (RLHF), fine-tuning with LoRA/QLoRA, and model evaluation benchmarking (MMLU, etc.).
- Study and implement Retrieval-Augmented Generation (RAG) systems end-to-end — he’s experimented but should productionize one.
- Contribute to a meaningful open-source LLM-adjacent project (e.g., llama.cpp, vLLM, or LangChain) — this is credibility gold for AI-centric job applications.
💼 3. Execute the Job Search Strategically
Zhiwei is open to AI Engineer, ML Engineer, Backend Engineer, Full-Stack Engineer, Staff Engineer, Tech Lead, and Founding Engineer roles. His preferred locations are Shenzhen/Guangzhou, with openness to Hong Kong and Singapore with visa sponsorship, and global remote teams.
Actions:
- Months 1–3: Polish LinkedIn, GitHub README (as a portfolio hub), and the English resume PDF. Apply to 10–15 targeted positions/month at AI-first companies and banks.
- Months 4–6: Prioritize Hong Kong and Singapore applications where compensation (40K HKD / 9K SGD/month) significantly exceeds current mainland package. Start visa sponsorship conversations early.
- Months 7–12: If no offer matches expectations, reassess — consider Founding Engineer or MTS roles at AI startups in Shenzhen/Guangzhou where equity could compensate for lower base.
🏗️ 4. Build One Flagship AI Product
Zhiwei has built tools like ww (CLI toolkit), iclaw (terminal AI agent), and zz (ML dataset utilities). His blog receives around 75,000 page views a month.
Actions:
- Pick one project (ideally
iclawor a new agent framework) and take it from personal tool → publicly usable product with documentation, a landing page, and GitHub stars. - Aim for 500+ GitHub stars on one project within 12 months — this signals real community traction to AI-centric hiring managers.
- Write 2–3 deep technical blog posts about the project (architecture decisions, benchmarks, lessons learned) to drive traffic and credibility.
🧪 5. Pursue a Recognized AI Certification or Research Output
Zhiwei has completed Machine Learning Specialization and Deep Learning Specialization from DeepLearning.AI/Stanford. He is a self-taught researcher with three academic-style papers on vision restoration.
Actions:
- Complete MLOps Specialization (DeepLearning.AI) or LLM Engineering courses (e.g., Hugging Face, fast.ai) to add credentials directly on the AI job search path.
- Write at least one computer science research-style blog post or preprint (e.g., on arXiv or his blog) about agent design, LLM prompt engineering, or small model training results. This bridges from vision-restoration researcher → CS researcher.
🌐 6. Improve English to Unlock Global Opportunities
Zhiwei’s IELTS score is 6.5 (Reading 8.5). He has taken online English courses with 60 Filipino English teachers. His compensation target in Hong Kong and Singapore requires strong professional English communication.
Actions:
- Retake IELTS aiming for 7.0+ overall (especially Speaking/Writing, likely the weak areas given Reading is 8.5). This is essential for HK/Singapore work visa applications and senior role interviews.
- Do 2–3 mock technical interviews in English per month (use platforms like Pramp, interviewing.io, or simply record yourself).
💰 7. Grow Financial Position (Side Track)
Zhiwei is a crypto and US stock investor since 2018, using long-term and regular investing strategies to achieve average annual returns of around 20%, though the absolute amount remains modest.
Actions:
- With an expected salary increase, systematically increase the monthly investment amount — the compounding effect is significant with a 20% average return.
- Maintain the discipline of not over-trading, which has been his edge so far.
🗓️ 12-Month Timeline Summary
| Period | Focus |
|---|---|
| Month 1–2 | Polish resume/LinkedIn, sit for 2–3 associate degree exams, apply to 15 AI roles |
| Month 3–4 | Sit for more exams, build flagship GitHub project, apply to HK/SG roles |
| Month 5–6 | Take IELTS retake, launch polished AI project publicly, write deep blog posts |
| Month 7–8 | Complete all associate degree exams, pursue MLOps or LLM certification |
| Month 9–10 | Write CS-domain paper or preprint, continue job search, assess offers |
| Month 11–12 | If offer secured, onboard and ramp up; if not, pivot to startup/founding engineer track |
The single most impactful thing: Finishing the associate degree removes a structural barrier, and landing a role where AI is the core job (not a side project) would align his daily work with where he is already investing all his personal time.
References: