Where I am useful
- Greenfield products. Take a product from first diagram to first release: domain model, API, frontend, deployment.
- Legacy systems. Map what exists, find the parts that hurt, and refactor in steps that keep the product running.
- Performance and scale. Slow APIs, heavy queries, infrastructure that struggles under load.
- AI systems. Agent workflows, private RAG pipelines, MCP servers, and the integration with the product you already have.
- Technical audit. An outside view of architecture, code quality, security and delivery risk, with a prioritised list of fixes.
Stack
| Area | Tools |
|---|---|
| Backend & core | Python (FastAPI, Django, Flask), Node.js, Java / Spring Boot |
| Frontend & web | JavaScript, TypeScript, React, Next.js, Astro |
| Data & cloud | PostgreSQL, MySQL, MongoDB, Redis, Docker, Kubernetes, AWS, Linux / VPS |
| AI & automation | Claude and OpenAI APIs, Ollama, LangChain / LangGraph, RAG, MCP |
| Product & delivery | Agile / Scrum, Jira, Confluence, roadmap and backlog work |
Expert level, in production for 15+ years: Python, JavaScript, Node.js. Advanced, 11+ years: React, TypeScript, MySQL. I use the rest when the project needs them.
The first two weeks
- Days 1–3: access, architecture walkthrough with the team, reading the code that matters.
- Days 4–7: first small change shipped through your pipeline, so I learn how delivery really works.
- Days 8–10: written assessment: what is solid, what is risky, what I would change first.
- From week 3: steady delivery inside your sprints, or the fixed scope we agreed.
Engagement models
| Model | What it looks like | Billing |
|---|---|---|
| Dedicated contracting | Full-time or part-time, inside your team’s sprints | Day rate or monthly |
| Fractional leadership | Fixed monthly hours for architecture decisions and reviews | Monthly retainer |
| Fixed-scope delivery | An MVP, a critical module or a refactor, end to end | Fixed budget per milestone |
Rates depend on duration and allocation. We set them on the first call.
Proof
- Products built end to end: espressimo, za•za•zu, Artelangelo.
- Technical writing: MCP tool calls vs. code execution, Private RAG reference architecture, Agent-ready APIs.