My Tech Pyramid
Technologies grouped by layer - design and direction at the top, runtime and data at the base.
Design & direction
- Solution architecture
- System design
- Product engineering
- UX & documentation
- API contracts
- Engineering leadership
- Applied AI workflows
Platforms & delivery
- React
- Next.js
- Vite
- Angular
- Laravel
- FastAPI
- Oracle ADF
- Docker
- Docker Compose
- AWS
- DigitalOcean
- Cloudflare
- nginx
- Caddy
- Apache
- PM2
- RabbitMQ
- Git
- GitHub
- GitHub Actions
- GitHub Pages
- Forgejo
- Linear
- Selenium
- MkDocs
Languages & application layer
- PHP
- Python
- TypeScript
- Java
- Node.js
- JavaScript
- Swift
- Objective-C
- C#
- REST
- OpenAPI
- Oracle MAF
Data, runtime & low level
- MySQL
- PostgreSQL
- Redis
- Supabase
- Oracle EBS
- PL/SQL
- MS SQL Server
- Linux
- Lean
- Typesense
- Solr
- JMeter
- C++
Impact areas
Strategic clarity - Make architecture, dependencies, constraints, ownership, and workflows legible for better investment and delivery decisions.
Root-cause diagnosis - Pinpoint failures under production pressure - from logs and routing through to the owning component and fix path.
Context switching - Move between domains, codebases, and time zones without cross-contamination - explicit boundaries and durable handoffs.
Dependable delivery - Connect diagnosis, narrow change, verification, and systemic learning so recovery strengthens the operating model.
Scalable platforms - Create direction, scaffolds, standards, and contribution paths so domain experts can own and extend systems.
Governed AI - Design AI workflows that increase judgment and execution speed while preserving evidence, boundaries, and human accountability.
Formal verification platforms - Turn Lean proofs, blueprints, and repository status into a product people can actually read - without inventing a green check that the code did not earn.
Engineering practice
- Docs-as-code - Markdown, MkDocs Material, GitHub Pages
- API contracts - design, endpoint validation, OpenAPI
- Controlled delivery - narrow branches, explicit promotion, runtime verification
- Cross-layer production diagnosis (logs, nginx, PM2, environment routing)
- Data migration planning and execution
- Web scraping, crawling, and integration APIs
- Legacy system maintenance, modernization, and performance tuning
- Visual system mapping and architecture synthesis
- Automated browser testing and UX verification
- Specialist agent roster - one workspace and skill file per client or stack
- Dual remotes - GitHub plus a self-hosted Forgejo mirror
- Secret hygiene - no credentials in docs, skills, or chat history
AI & automation
- Local-first LLM workflows (LM Studio, MLX)
- Checkpointed background workers and resumable pipelines
- Deterministic preprocessing and bounded model output
- Cursor-assisted development and agent orchestration design
- Privacy-tier design - employee filtering, collection boundaries, immutable sources
- Structured Markdown knowledge stores and Tier-1/Tier-2 synthesis
- Semantic event preprocessing and task-fit model routing
- Code computes, models narrate - grades, weights, and counts stay in deterministic code
- Self-learning playbooks - a repeated mistake updates one file, not a longer prompt
Leadership & management
- Engineering manager / lead - teams to ~15 engineers; delivery, stakeholder and shareholder reporting
- Solution architecture - system design, integration boundaries, platform direction across services and repos
- Global programme delivery - discovery through production, on-site and remote
- Technical direction across services, repositories, and environments
- Mentoring through context - scaffolds that enable ownership
- Executive translation of architecture, incidents, and trade-offs
- Operational credibility - close enough to code and production to verify standards
- Recruitment process design and team coordination