Space 03/Mind
Capabilities for the AI era.
Frameworks change every year. What compounds is knowing what to build, how it should scale and how to ship it — with AI as leverage, not a crutch.
Why the list looks like this
Codeisgettingcheapereverymonth.Whatdoesn'tgetcheaperisknowingwhattobuild,designingsystemsthatsurvivesuccess,andshippingsomethingpeopleactuallyuse.
AI-native building
Leverage, not autocomplete.
Designing products where models do real work — and engineering the guardrails that make that work trustworthy.
- 01.1
Agentic workflows
Breaking a job into tools, steps and checks an agent can actually finish, with a human in the loop where it counts.
- 01.2
Context & prompt engineering
Treating prompts and retrieved context as versioned product code, not magic strings.
- 01.3
RAG & grounding
Answers that come from your data instead of model memory, with sources you can check.
- 01.4
Evals & reliability
Test sets, failure tracking and fallbacks, so quality is measured rather than assumed.
- 01.5
Voice & real-time AI UX
Speech interfaces that feel conversational: streaming, interruptions and tight latency budgets.
- 01.6
AI-assisted engineering
Using coding agents to move faster while staying the one accountable for architecture and review.
Product & startup building
Zero to one, then one to ten.
Turning a fuzzy idea into the smallest product worth shipping, then improving it on real usage.
- 02.1
Problem framing
Finding the real job to be done before writing the first line of code.
- 02.2
MVP scoping
Cutting scope to what proves the idea, and saying “not yet” to the rest.
- 02.3
Rapid prototyping
Working versions in days, so decisions get made on evidence instead of slides.
- 02.4
Product sense & UX
Caring how it feels to use, not just whether it compiles.
- 02.5
Metrics & iteration
Instrumenting launches, reading what users actually do and shipping the next improvement.
- 02.6
Shipping velocity
Small releases and short feedback loops instead of big-bang launches.
System design & scale
Built to survive success.
Architecture that holds up as users, data and teams grow — without over-engineering day one.
- 03.1
Architecture & trade-offs
Boring technology where possible, complexity only where it pays for itself.
- 03.2
Data modelling & API design
Schemas and contracts that make the next feature easy instead of painful.
- 03.3
Multi-tenant SaaS
Tenant isolation, roles and permissions designed in from the start.
- 03.4
Real-time systems
WebSockets and streaming pipelines for live, low-latency experiences.
- 03.5
Security & privacy by design
Healthcare-grade habits: least privilege, scoped access and audit trails.
- 03.6
Reliability & cost awareness
Monitoring, graceful failure and infrastructure that doesn't burn the runway.
Business & leadership
Engineering that speaks business.
Translating between what the business needs and what engineering can deliver — clearly, early and honestly.
- 04.1
Client & stakeholder communication
Working directly with clients on scoping, demos and the hard conversations about trade-offs.
- 04.2
Estimation & delivery
Realistic plans, visible progress and no surprise deadlines.
- 04.3
Technical writing
Specs, docs and handovers that let someone else pick the work up.
- 04.4
Teaching & mentoring
Explaining hard ideas simply — practised every week teaching kids and teenagers to code.
- 04.5
Ownership
Owning the outcome end to end, not just closing the ticket.
- 04.6
Learning velocity
Picking up new tools fast, which matters when the stack changes every quarter.
Engineering craft
The foundation everything stands on.
The hands-on stack I ship with every day — chosen for speed without giving up control.
Frontend
- React
- Next.js
- TypeScript
- Tailwind CSS
- React Native
- Expo
Backend
- Node.js
- Express
- FastAPI
- Python
- REST APIs
- WebSockets
Data & Platform
- Firebase
- PostgreSQL
- MongoDB
- Vercel
- Git
- CI/CD
AI Engineering
- LLM integration
- Voice interfaces
- Prompt design
- RAG
- AI agents