Tools & Platforms
The working stack. Tools are means to an end, and I use what fits the team already in place.
AI and Agentic Tooling
- Claude Code: daily, for building and review
- Claude API and MCP: agent and tool integrations
- Agent skills: instruction files that teach an agent a team's conventions
- RAG content design: self-contained, intent-tagged sections and Q&A pairs, chunked for retrieval
- LLM-advisory architecture: the model advises and vetoes, deterministic code executes
I read and review code; I ship with agentic tooling.
Delivery and Programme
- Jira and Confluence: delivery process, Kanban and Agile boards, documentation
- Figma: content design on product screens
- Mermaid: architecture diagrams that live in version control
Engineering and Infrastructure (working level)
- Languages: PHP, JavaScript and TypeScript, Node and Bun, SQL (MySQL, Postgres), HTML, XML, JSON
- APIs: REST and webhooks
- Version control and CI/CD: Git, GitHub Actions, GitLab CI, Jenkins
- Cloud: AWS (EC2, S3, ELB, Auto Scaling, RDS), Docker
- Caching and edge: Redis, Memcached, Varnish, Cloudflare CDN and WAF
- Observability and performance: New Relic, Blackfire, Lighthouse
- Linux CLI
Commerce
- Platforms: Magento 1 and 2 (Enterprise, Commerce Cloud), Shopify, BigCommerce
- Headless: Vue, React, GraphQL
- PIM and search: Akeneo, Klevu, Elasticsearch, Solr
- ERP: Microsoft Dynamics NAV, NetSuite, Sage, SAP
Payments and Blockchain
- Card tokenisation: Mastercard MDES and MDES Manager
- Digital wallets: Apple Pay, Google Pay and Samsung Pay
- PSP and gateway integration: Sage Pay, WorldPay, PayPal, Klarna, Amazon Pay
- Ethereum Layer 2: Starknet, Cairo and STARK proofs, documented at reading level
- On-chain analysis: Solana, Ethereum, GMGN, Cielo, Solscan, Bubblemaps
Content and Documentation
Enterprise structured authoring
- DITA XML and DITA-OT: structured authoring and a local build toolchain at Mastercard
Docs-as-code
- Docusaurus: Starknet
- Nextra: this site
- MkDocs, GitBook, AsciiDoc
- Markdown and MDX
Quality
- Vale: prose linting
- W3C and COGA accessibility guidelines
Ramp-up Time
Productive within:
- 1 day: an existing delivery or docs toolchain
- 1 week: a new codebase, enough to review architecture and make calls
- 2 to 4 weeks: a genuinely new technical domain, working with the people who know it
I will not claim instant expertise in something I have never seen.