Gina Horch

I'm a software developer working across TypeScript and Python, building full-stack production systems with an AI-augmented engineering practice — context, constraints, review and verification, not just prompting and accepting. Day to day that means React and Next.js applications, an Astro site in commercial production, PostgreSQL and Supabase behind them, Python-based automation and integrations, and the webhooks and scheduled jobs that connect those systems to everything else an organisation already runs on.

The work I care most about is the kind that has real users attached to it: rebuilding a company's public website and planning its production migration; taking over a partially built product and getting it into a reliable state; developing and maintaining a data platform where a wrong assumption about identity can quietly corrupt years of history. Building something new is enjoyable. Understanding a system you did not write, and changing it safely, is the harder and more useful skill.

I work with AI-assisted development tools - Claude Code, Cursor, Replit - as part of a disciplined engineering workflow rather than as a shortcut. That means persistent project context and constraints, planning before implementation, reviewing what comes back, testing behaviour, diagnosing failures, and keeping a clear Definition of Done. I remain responsible for the requirements, the technical decisions, the security considerations and the finished system.

Security and systems thinking run through all of it. I work on authentication and access control, Row Level Security, privacy and consent, secure configuration and security headers, and I hold a Diploma of Information Technology specialising in cyber security and business analysis. Treating security as part of engineering, rather than a pass to run after the build, is considerably cheaper and produces better software.

I came to development along an unusual path: twenty-two years in law enforcement leading complex investigations, then research with the Australian Centre for Child Protection, then technology. Digital forensics is where I first saw how much technical work sits underneath a good answer, and research is where I saw how many processes could be improved by building something better. That background is why I approach problems the way I do — establish what is actually true before acting, look for the evidence rather than the assumption, document what you found, and think carefully about the people whose data you are handling.

I work collaboratively and iteratively: short feedback loops, transparency about what is and isn't done, and reviewing outcomes rather than assuming that shipping equals finished. I'm based on Whadjuk Noongar boodja in Perth, Western Australia.