Full-stack engineer focused on backend architecture, data systems, and AI applications.
Hi, I'm Paolo. I build reliable systems that connect robust backend architecture with intuitive frontend experiences. I work across scalable APIs, data workflows, application-layer AI engineering, and production delivery.
- Backend services and REST APIs with clear boundaries, dependable data flows, and practical observability.
- AI and RAG applications that keep outputs grounded, inspectable, and useful to the people making decisions.
- Automation workflows that connect Gmail, Google Sheets, Google Calendar, and business processes.
- Production-ready web applications with responsive interfaces, Playwright coverage, and end-to-end testing.
I focus on practical AI engineering at the application layer: connecting transformer-based models to reliable software systems instead of training foundation models from scratch.
- Autonomous AI agent workflows: building bounded, tool-using workflows that classify work, call APIs, keep task state, and route consequential actions through explicit human approval.
- Transformer architecture in practice: working with model context, embeddings, prompting, structured outputs, and API orchestration to build useful features around transformer-based LLMs.
- Persistent memory and retrieval: using bounded chat memory, PostgreSQL/pgvector, vector search, semantic caching, and RAG to keep context useful, scoped, and inspectable.
- Reliable AI delivery: adding evaluation sets, prompt-injection checks, idempotent workflows, guardrails, and clear failure paths around AI features.
| Project | Focus | Link |
|---|---|---|
| Mailroom | Private AI email workflows with Gmail triage, grounded drafts, and calendar automation. | Private project; details available on request |
| MTG Rules Desk | Citation-first RAG for Magic: The Gathering rules, Oracle text, and rulings. | Live app |
| MarketView | Market data collection, analysis, and visualization for data-driven decisions. | Live app |
Explore the full project archive in the portfolio.
| Category | Stack |
|---|---|
| Languages | C++ · C# · Python · JavaScript · TypeScript · PHP · SQL · HTML/CSS |
| Frameworks & Libraries | React · Vite · FastAPI · REST APIs · Pandas · NumPy · Matplotlib · Tauri 2 · RAG |
| AI Engineering | Autonomous AI agent workflows · Transformer-based LLMs · Tool calling · RAG · Embeddings · Persistent memory · pgvector · Semantic caching · Human-in-the-loop safety |
| Databases | SQL · PostgreSQL · Firebase · Supabase · Upstash Redis · Upstash Vector · pgvector · MongoDB |
| Tools & Platforms | Docker · Git · GitHub · Google Cloud Platform · n8n · OpenAI · Playwright · Vercel · Terraform · Hostinger · Dokploy |
| Soft Skills | Adaptability · Problem Solving · Teamwork · Work Ethic |
- Make systems understandable: document decisions, keep boundaries explicit, and prefer simple flows that are easy to inspect.
- Keep humans in control: design AI and automation with safe defaults, authorization checkpoints, and grounded outputs.
- Ship with confidence: use focused tests, end-to-end checks, and repeatable delivery workflows before calling work complete.
Building useful systems, one careful boundary at a time.