Qiwei He

Role: Software engineer, and interim team lead
Stack: React · TypeScript · Tailwind CSS · Material UI · C# · Claude Code · Cursor

Built and maintained production frontend applications for a large enterprise services business, and helped change how the team shipped. Client name withheld.

Delivery

• Developed and maintained scalable, responsive frontend applications in React and TypeScript with Tailwind CSS and Material UI
• Implemented Backend-for-Frontend services in C# to streamline data integration between the client and downstream APIs, measurably improving client-side performance
• Built custom CMS endpoints to improve data retrieval efficiency and cut redundant API calls
• Acted as interim team lead in the lead’s absence, coordinating sprint work and managing production deployments
• Carried on-call duty for production reliability and incident response

AI-assisted engineering

• Integrated Claude Code and Cursor into daily development, automating test-case generation, pull-request creation and code review
• Authored project-specific Claude Code skills and rule sets that were adopted across the engineering team, standardising how AI-assisted development was done rather than leaving it to individual habit

That last point is the part worth hiring for. Plenty of engineers use AI tools; comparatively few have had to decide what good practice looks like for a whole team and then write it down.

Role: Front-end team lead — architecture, standards and hands-on delivery
Stack: React · TypeScript · Redux · Three.js · Python · C#

Led the frontend of a commercial SaaS platform for the insurance industry, from system architecture through to production delivery, in a small hands-on team. Client name withheld.

What I owned

• Designed the frontend system architecture and established the component patterns, coding standards and review practices the team worked to
• Primary developer on a browser-based PDF annotation tool that was a core feature of the product suite
• Built 3D infrastructure viewers in Three.js capable of rendering and interacting with over 25,000 IoT devices in a single scene
• Built a multi-file upload system for the platform’s CMS, improving content workflows for internal teams
• Developed an automated pipeline in Python and C# that converted CAD drawings into AI training data, supporting the company’s machine-learning product
• Mentored junior developers and interns through code review and technical guidance

The 3D and document-annotation work is the unusual part: both required getting heavy third-party rendering libraries to perform well inside a React application, at data volumes that break naive implementations.

Role: Sole engineer — backend, cloud infrastructure and release pipeline
Stack: NestJS · TypeScript · PostgreSQL · TypeORM · AWS · Terraform · Terragrunt · GitHub Actions

Built the production backend for a mobile camera-accessory platform end to end, and own everything underneath it. Ongoing engagement; client name withheld.

What I built

• REST API in NestJS and TypeScript on PostgreSQL, with OpenAPI documentation generated from the code via Swagger
• Authentication and account management: JWT access and refresh token rotation, password reset and email verification, with transactional email delivered through SendGrid
• Apple App Attest integration, so the API can verify that requests originate from genuine instances of the iOS app
• Apple Universal Links, so password-reset and verification emails deep-link straight into the app
• Product ownership and warranty registration APIs covering document upload and review workflows
• Every schema change shipped as a reviewed TypeORM migration

Infrastructure and delivery

• The entire AWS footprint written as code in Terraform and Terragrunt — ECS Fargate, Application Load Balancer, RDS PostgreSQL, Secrets Manager and ECR — across fully isolated staging and production environments
• GitHub Actions CI/CD on keyless OIDC authentication, with automated staging deploys and a manual approval gate on production
• An AWS SageMaker training pipeline and S3 dataset structure supporting an on-device product classification model