About

Updated on: May 2026

Email: [email protected]

About This Blog

• Reading leisurely books, daydreaming, judging people at a glance, and making snide remarks.
• Often imaginative, occasionally contemplative.
• Taurus sun, Aquarius moon, Scorpio rising, with many planets in the 3rd house.
• May the world find peace and good health.

About Me

I'm someone who continuously integrates experience and works to match my abilities with real-world needs. I don't follow a conventional career trajectory — instead, I have gradually accumulated skills and refined my direction through practice, realising along the way:

Not everyone can readily conform to a specific job title, but everyone has value to offer.

Transitioning between niche fields, I have never found cross-disciplinary moves particularly troubling. Each shift was a deliberate choice made after discerning (identifying) my abilities and interests at that stage. My exam performance is average, but I have never worried about failing a course. I consistently ranked just below the scholarship cutoff, and I seem to have an aptitude for research.

I am seeking a mode of work — it need not be a dream position, but I hope to collaborate with organisations that are patient, supportive, and respectful of realistic pacing (steady rhythms). Although I have long operated in a "peripheral research role," as long as the structure is stable and I can perform reliably, I am willing to start from the ground up.

If you are looking for someone who:

• Is technically capable but not obstinate; understands boundaries but does not overstep;
• Can transform disparate materials into polished deliverables;
• Will not overexert but will not evade responsibility; can handle medium-to-high intensity tasks autonomously;
• Is responsive to external changes without being anxious; adapts well to steady progress;

Maybe I'm a suitable collaborator.

Quanzhou Normal University — B.Eng. in Marine Technology

September 2015 – June 2019

Mastery of fundamental theories, knowledge, and skills in marine biotechnology and marine science; ability to apply advanced marine biotechnology for research and development; awareness of innovation and entrepreneurship; capacity for lifelong independent learning; adaptability to both independent and team work environments. Graduates are equipped for scientific research, teaching, management, and technology development in areas such as marine bioresource survey and utilization, aquaculture, marine environmental monitoring and management, and new marine technology R&D. (Quoted from the Program Educational Objectives.)

Both faculty and students made genuine efforts to meet these objectives, yet societal demand for marine biology graduates remains limited, and the field is not widely acknowledged.

After completing my first year, I still had little understanding of what I was meant to learn or what our major entailed. By chance, I became an apprentice to Mr. Tsen Suh-der from National Yang Ming University, raising Japanese eels under his guidance.

Raising eels demands a firm grasp of animal physiology — it involves dissection, injections, surgery, tissue sectioning, and cell observation. After two years of wet-lab work, the project came to an absurd end when the department stopped providing lab space to Mr. Tsen.

Functional genomics was a prominent research direction at the time. My graduation project shifted toward bioinformatics, which became my gateway to programming. The life sciences did not miss the "big data" wave — high-throughput sequencing was gaining momentum. After experimenting with machine learning, I did not jump directly into "AI" but instead focused on transcriptomics for my thesis.

• Transcriptome analysis of Anguilla japonica under salinity adaptation
• Work: Led internal group discussions on analysis pipelines and result interpretation, and drafted the final paper.
  • Simulated oceanic migration conditions by gradually increasing salinity in culture tanks.
  • Collected RNA-seq samples from brain, gonad, and pectoral fin tissues.
  • Conducted downstream transcriptome analysis: differential expression, gene function annotation, and phylogenetic tree construction.

Looking back, the undergraduate training I received was more than adequate for a 211-programme master's degree.

March 2017 – July 2018

I unfortunately took on a side project that yielded no tangible outcomes. We were a cross-department entrepreneurial team. After graduation, the successful members went straight to CTO roles at startups. As the only person in the tech department alongside the star player, I was essentially carried (if you will pardon the self-deprecation).

• Technology promotion: Organised project roadshows and tech salons to promote cutting-edge technologies such as 3D printing, drones, and cloud computing. Researched and practised mainstream FDM 3D printing, independently resolving common structural issues like warping and layer lines.
• School-enterprise collaboration: Assisted in building and promoting a student entrepreneurship incubation base, coordinating team communication, follow-ups, and feedback collection. Conducted surveys to gather and analyse operational data.

Without this entrepreneurial detour, I might have graduated as a dullard who knew little more than "life sciences" blended with "marine science." CET-6, the National Computer Rank Examination Level 2, the postgraduate entrance exam — the trifecta of mundane undergraduate achievements, maxed out.

Shantou University — Guangdong Key Laboratory of Marine Biotechnology

June 2019 – July 2021

I applied to a university near my hometown for graduate school — ranked 3rd in the entrance exam, and started working right after graduation.

By the time the semester was about to begin, I could no longer tolerate the toxic PI situation and withdrew (dropped out).

• Publication: Frontiers in Marine Science, 2021
• Project: Constructed a full-length transcriptome of male and female gonadal development in the mud crab Scylla paramamosain using PacBio SMRT and Illumina sequencing.
• Work: Bioinformatics, experimental design, and data analysis
  • Conducted downstream analysis: alternative splicing and differential gene expression (DEG) identification.
  • Designed and executed qPCR and RT-PCR experiments to validate sequencing results.
  • Responsible for RNA extraction and sample preparation.
  • Contributed to data visualization and manuscript drafting.

I handed in a draft manuscript. Regrettably, my labmates were wonderful people — I did not get the chance to become close friends with them.

BTW, after I left: 1 master's dropout, 2 PhD dropouts, 1 resigned research assistant, 2 postdoc resignations, and 3 master's students with delayed graduations.

Due to force majeure, I abandoned my plans to study abroad or in Taiwan.

My odd-job period did not last long — I did not want a disconcerting thought to become reality:

"Is dissecting crabs the closest I will ever get to being a graduate student?"

So I took and passed the exam again. Ranked 2nd in the entrance exam.

Huazhong Agricultural University — Imperial College London Summer School: Computational Social Science Methodology

July 2021

The summer school was my first assignment after being admitted. The final project required building a small model in NetLogo.

• Agent-based SIR epidemic transmission and vaccine intervention simulation system
• Project: Added vaccine intervention to the SIR model, simulating behavior and interactions in a social network, evaluating vaccine effectiveness at different coverage rates, and predicting epidemic trends to inform COVID-19 vaccine policy.
• Work:
  • Discussed and finalized model design details with the team (agent characteristics, environmental factors, interaction rules);
  • Implemented the model in code, presented results through charts and graphs, and wrote model documentation;
  • Assisted teammates in collecting, organizing, and calibrating relevant parameters;
  • Conducted model validation and local structural sensitivity analysis.

Another carried team project — effortlessly bringing along a top-2 university teammate and a PhD student of the same age.

Complex adaptive systems reshaped my understanding of biology. I was no stranger to multi-agent system modelling — I had worked on a similar (though unremarkable) project in 2020. Moving from an intuitive grasp to a rigorous understanding, it felt inevitable that this would become my research direction.

Liaoning University — M.Sc. in Biostatistics

September 2021 – June 2024

Knowledge structure reasonably sound, with a certain keen insight, innovation ability, and academic research capability; skilled at combining biology, statistics theory, and practice; able to independently propose, analyze, and solve problems; adaptable as a specialized talent in bioinformatics and biostatistics for societal needs. (Also quoted from the Program Educational Objectives)

I regret that I did not achieve a "reasonably sound knowledge structure," but the university's lax adherence to its own training objectives was beyond student control. As for whether graduates can adapt to society — in the current climate, assigning blame on either side is futile. Regarding other competencies, I am fortunate to have received solid training during my undergraduate years.

A disastrous start — the very first group meeting revealed the first delayed graduation case in our advisor's group. Fortunately, history did not repeat itself; the path was merely more arduous. Although my advisor was not particularly competent (somewhat lacking in responsibility), he was at least a decent person.

The advisor was hands-off, and we became slaves to that freedom. "Reasonably sound knowledge structure" became an elusive goal.

Every student worked on an independent project. My first choice was causal inference. I suppose I took some detours during my first year. I joined a team competition after a long hiatus — the project was chosen by my supervisor to ride the "carbon neutrality" wave. Expecting elimination in the first round, our model unexpectedly made the cut, forcing me to wake up early on a summer weekend for an online presentation.

March 2022 – September 2022

• Plant carbon sequestration potential assessment model based on population life history
• Project: Studied native elm woodland in Otindag Sandy Land and Horqin Sandy Land, collected plant growth data, extracted regional precipitation from Shapefile data. Built a system dynamics model simulating plant growth from seed to mature tree under different precipitation scenarios, evaluating the model's accuracy and application value.
• Result: Model was shortlisted for the China Simulation Federation Complex System Simulation Competition.

Because of the extended lockdowns in 2022 that lasted months, I had ample time for reflection at home. To secure graduation, I took over a legacy project that had been abandoned by two cohorts of students — an agent-based evolutionary game model in ecology, bringing me back to the complex systems track.

Theoretical biology is a very niche direction. My proposal defense was met with silence — no one opposed, but only one instructor (now our programme director) believed I could see it through. It is not that I felt no pressure. A week later, ChatGPT was released, and I could already sense that many of my experiences would soon become obsolete.

The legacy project? Finished in one winter break. I spent over six months deliberately feigning normal productivity in my progress reports to my supervisor, occasionally adding finishing touches.

Two thoughts for 2023:

1. Confirm whether I still need a PhD (or rather: whether I can still consider doing a PhD in China)
2. Spend more time exploring the possibilities of LLMs.

After the semester started, I did not return to campus. Instead, I accepted an invitation from a postdoc from my former lab to visit his institute.

Southern Marine Science and Engineering Guangdong Laboratory (Zhanjiang)

April 2023 – May 2023

My senior colleague had been researching seaweed molecular evolution and was applying for funding to build a local marine algae resource database for data-driven research. The team was just forming, having recently obtained joint graduate training qualifications with local universities. He wanted graduate students to conduct experimental projects and invited me to share my experience.

• Data management: Participated in discussing and drafting grant proposals, designed database templates, improved lab data collection efficiency, ensured data standardisation and integrity.
• Teaching: Leveraged my communication skills to help biology graduate students quickly master simulation, statistical modelling, and machine learning — covering both relevant theory and software operation.

We spent about two months together — collecting intertidal samples, then hiking 9 km through a village. A genuine camaraderie forged under pressure. But witnessing how academician-level groups absorbed young scholars' funding nearly extinguished my plan to remain in the domestic academic sphere.

Both of my questions were partially answered.

I delegated AI-related matters to AI tools, deployed a machine learning workflow, confirmed that the students had mastered the fundamentals, and returned to school.

University of Aberdeen — Overseas Distinguished Scholar Short-term Course: Advanced Econometrics

June 2023

I sat in on the previous cohort's thesis defence and saw my advisor — who had just announced his departure. I submitted a partial Chinese manuscript to him, hoping to meet the minimum graduation requirements as soon as possible. His reply was predictable: "Translate it into English."

Around the same time, the economics department was hosting various short courses led by overseas distinguished scholars. I enrolled in Advanced Econometrics and, together with a physics major and an economics major, ranked first in the class. For the final project, I selected Flammer (2021)'s paper on Corporate Green Bonds and wrote a critique.

Agricultural Bank of China Summer Internship

July 2023 – August 2023

Understanding bank structure and core business: deposits, loans, remittances. Learning customer communication skills: handling inquiries and complaints, improving service awareness. Mastering features, target customers, and risk control of major products: credit cards, personal loans, wealth management products. Learning compliance knowledge: anti-money laundering, anti-fraud. Identifying and preventing financial risks, building risk awareness.

The bank internship was intellectually unfulfilling. In my spare moments, watching the recruitment season arrive earlier each year, I mechanically registered accounts, filled in forms, and submitted applications — yet received no feedback by what should have been peak recruitment season.

Then came the mid-term review. Professors who had remained silent during my proposal suddenly turned hostile and suggested I discuss changing my topic with my advisor. But my supervisor had already left — he was not there. The atmosphere was tense. Changing topics would mean an immediate half-year delay. The only way out: get my paper accepted by an SCI journal.

Right at that moment, I received exam and interview notifications from several organisations. The only offer was an internship as a SAS programmer.

It offered frequent work-from-home opportunities, and at the time, it was the most concrete, straightforward career path I could envision — processing clinical data with code, clear tasks, no need to be a jack-of-all-trades. Of course, it is probably a sunset profession.

I then incorporated Sobol' global sensitivity analysis to strengthen my paper. After 2–3 months of submissions (rejected by three different journals), my supervisor still had not reviewed it once. I grew accustomed to emailing unfamiliar professors directly for help. After much scrambling, I missed the internship start date.

The pre-defence (preliminary oral defence) arrived. The department was still "concerned" that my thesis would not pass blind review.

May 2024

A nerve-wracking six months. I wanted to be confident but had no benchmark to gauge myself against. After completing the thesis, I proofread it with my former competition teammate and submitted it for blind review.

I only received the blind review results two days before the defence — mediocre, supposedly ranked 2nd.

Yet someone on the defence committee still abstained from voting.

Present

September 2024 – December 2024

State-owned enterprise offer → medical check → position cancelled.

2025

GCP certification → abandoned the CRO (Contract Research Organisation) industry.

I currently work in independent research services: providing experimental design, public database mining, data analysis, paper submission, and revision support across seven different fields (biology & ecology, clinical & pharmaceutical, economics & statistics, civil engineering, etc.). I also offer development services within my capability, such as small websites, web applications, crawlers, and data visualisation dashboards.

Interests

• Cutting-Edge Biological Research
  • Medical Field
  • Marine Science
  • AIVC (Artificial Intelligence Virtual Cell)
• Theoretical Biology and Evolutionary Biology
  • Complex Adaptive Systems
• Agentic Engineering
  • Game Development (Bevy)
  • Building Something...
• Language Learning
  • Japanese (N3)
  • German (A2, grammar only)

Maintained Projects

Multi-Platform Cross-Border E-commerce Automated Monitoring & Price Comparison System | April 2025 – Present

An LLM + Playwright based intelligent crawler and data analysis system for automated collection, cleaning, and analysis of price, inventory, and review data across multiple cross-border e-commerce platforms.

• Uses LLM to generate Playwright scripts for target pages, with built-in request header spoofing, proxy rotation, and exception retry; scripts return structured JSON, avoiding the high token consumption of LLM + MCP direct browser control.
• Combines OpenClaw Skill + CLI to manage context overhead, enabling the agent to autonomously handle CAPTCHAs, cookie expiration, API rate limiting, and other edge cases.
• Builds XGBoost regression and Prophet time series models on historical price data for trend prediction, arbitrage opportunity identification, and anomaly detection.
• Modular architecture separates the crawler engine, data processing, anti-crawl strategy, and frontend display, supporting independent scaling and distributed deployment.
• Provides data visualisation including price trend charts, inventory alerts, and competitive analysis.