Jinping HeSenior Full-Stack Developer & Data Analyst · Auckland

Blogs

Notes on design, delivery, and the systems behind them.

Insights · 7 Jul 2026

.NET vs Java Microservices in New Zealand: A Practical Guide

For New Zealand companies adopting microservice architecture, choosing between .NET and Java depends on ecosystem alignment, cloud strategy, talent availability, and long-term enterprise requirements.

Insights · 12 Aug 2026

Microservices Performance from a Full-Stack View: Lessons from a SaaS

A practical review of performance decisions in a multi-tenant online assessment SaaS, covering the gateway, authentication, data access, frontend, and judging flow. Useful for developers working with .NET, Next.js, and microservices

Insights · 25 Jul 2026

From Power BI to a Full-Stack Loan Pre-Check Demo

Starting from a Power BI assignment on the Home Credit Default Risk dataset, I extended the project into a multilingual full-stack machine learning demo called HomeCredit Scout. This article shares the journey from data analysis to product prototyping using LightGBM, FastAPI, PostgreSQL, React, and practical considerations such as privacy and multilingual UX.

Insights · 7 Jul 2026

Power BI vs Python: Choosing the Right Data Analytics Tool

Power BI and Python solve different data challenges. Understanding when to use business intelligence dashboards versus machine learning models is critical for making better data-driven decisions.

Insights · 7 Jul 2026

Monolith vs Headless: Choosing Architecture Based on Business Needs

Many developers assume that frontend-backend separation is inherently more modern than a monolithic architecture. In reality, the right choice depends on business requirements, team size, budget, and long-term maintenance goals—not trends or technical fashion.

Insights · 7 Jul 2026

China's Tutoring Industry After Double Reduction

Nearly five years after China introduced the Double Reduction Policy, the country's tutoring industry looks very different from the one that once dominated the education market. While traditional K–12 academic tutoring has largely disappeared from public view, the sector has not vanished. Instead, it has undergone a significant transformation, with companies shifting toward quality education, vocational training, study-abroad services, and technology-driven learning solutions. This article examines the current state of China's tutoring industry, the major trends shaping its future, and what education entrepreneurs can learn from this period of rapid change.

Insights · 7 Jul 2026

NZQA’s 2026 PTE Rule Changes: A Win for New Training Providers?

New Zealand’s new Private Training Establishment (PTE) Rules 2026 reduce some long-standing compliance burdens by removing annual fees and cutting financial reporting requirements from yearly to biennial submissions for many providers. For education entrepreneurs, the changes create opportunities to lower administrative costs and simplify operations. However, reduced reporting does not necessarily mean reduced oversight. Here's what the new rules mean for anyone considering launching or scaling a PTE in New Zealand.

Insights · 7 Jul 2026

Why New Zealand PTEs Are Betting on Data, Education, and Healthcare in 2026

As New Zealand's workforce shortages continue to drive change, Private Training Establishments (PTEs) are increasingly focusing on high-demand fields such as Business Intelligence, Early Childhood Education, and healthcare. This trend reflects New Zealand's employment-driven approach to education and highlights the growing importance of skills that align with labor market needs.

Insights · 7 Jul 2026

Cursor vs. Codex: A Personal, Unscientific Take After Actually Using Both

A subjective, no-sponsorship comparison of Cursor and Codex based on real day-to-day use — not a benchmark, not a verdict, just one developer's honest impressions of what felt natural and what didn't.

Insights · 23 Apr 2026

Which Jobs Are Most Vulnerable in the Age of AI?

AI is transforming the workforce faster than many expected. The jobs most at risk are not necessarily the lowest paid, but those built around repetitive information processing and predictable decision-making.