Software Engineer focused on backend systems, production reliability, cloud deployment, and AI-assisted workflow automation.
I'm a software engineer in Sydney building practical systems around real workflows: production recovery, CI/CD, cloud deployment, internal tools, and AI-assisted automation. My strongest work sits at the intersection of shipping features, debugging production issues, and making systems easier to operate.
Recent experience includes production recovery and marketplace platform work at Ask Jay Services, where I helped restore service functionality after a security incident, improved load times from roughly 25 seconds to under 3 seconds, and built Docker + GitHub Actions deployment workflows.
I also build prototypes and open-source tools to explore responsible AI workflows, developer tooling, and small-business automation. I label these clearly as MVPs, prototypes, or experiments when they are not production systems.
Languages & Frameworks:
Cloud & DevOps:
Databases & Storage:
Specialties & Methodologies:
- Ask Jay Services: Founding Engineer / Principal Technical Lead, May-Aug 2025. Production recovery and marketplace platform work, including service restoration, load-time improvement from roughly 25 seconds to under 3 seconds, and Docker + GitHub Actions deployment workflows.
- Australian Computer Society: Web Developer Intern, Sep 2023-Feb 2024. Worked on a MERN platform serving 10,000+ users, improved average page load time by 30%, and reviewed/resolved 15+ authentication issues using OWASP guidance.
The portfolio separates flagship case studies from the broader GitHub project feed:
- Production Recovery & Performance Rebuild - NDA-safe case study for production recovery, cloud hardening, CI/CD, and performance work.
- Pilly / MediMate Voice - Firebase-backed responsible-AI medication support MVP. Not a medical product; no diagnosis, dosage advice, or real patient data.
- codeflow-hook - Open-source AI-assisted code review CLI published as an npm package with early usage traction.
The website project feed is automated from public, non-fork GitHub repositories owned by Sharv619. By default it uses PORTFOLIO_GITHUB_TOPIC=all; set this to a specific topic such as portfolio if the feed should be curated. Keep repo descriptions/READMEs useful, and the next GitHub Pages rebuild will refresh the portfolio pages. The deploy workflow also runs on a daily schedule so project metadata updates without manually editing src/lib/data.ts. Set the GitHub Actions variable NEXT_PUBLIC_ASSISTANT_API to the RAG Lambda endpoint when the deployed chatbot should use the live GitHub-backed assistant instead of demo mode.
Automation details for repo push refreshes and certifications live in docs/automation.md.
I'm always open to connecting with fellow builders and exploring new opportunities:
