- 🧩 I build production-shaped backends — opinionated Django foundations, identity/auth engines and third-party API integrations, taken from first commit through to a containerised deployment.
- 🔐 My strongest ground is security-first architecture: JWT rotation and blacklisting, MFA over TOTP, application-level encryption, brute-force lockout, structured audit logging.
- 🤖 I also design AI agent governance tooling — zero-trust subagent pipelines with hard token budgets, shipping as Token-Optimized-Agent-Pipeline.
- 🎓 Self-taught, through courses and by shipping production-style projects end to end. Based in Madrid and open to backend roles.
| Languages |   |
| Backend |    |
| Data |   |
| DevOps & Tooling |     |
| Project | What it is | Stack |
|---|---|---|
| Token-Optimized-Agent-Pipeline | Zero-trust subagent architecture for token-efficient AI software engineering — 34 skills, 13 agents, 11 slash commands | Python |
| Django-Users-App | Installable Django app for identity and access management — JWT, TOTP, step-up auth, encrypted PII | Django · DRF · PostgreSQL |
| Django-Pro-Template | Production-ready Django foundation with a secure, scalable architecture | Django · Docker |
| Tradingview2EXCH | REST API bridging TradingView webhook signals to the Binance exchange | Django · DRF · Docker |
🤖 Token-Optimized-Agent-Pipeline — AI agent governance framework
An AI agent governance framework that turns general-purpose agents into a coordinated pipeline of specialists. Applies the same rigour to itself as to every host it audits.
- Governance ruleset constraining agent behaviour, code quality and project topology.
- Eight-phase pipeline (Planning → Sprint Closeout) gated by a single attended human authorisation.
- Specialised subagent roles — orchestrator, rule validator, QA, tester — instead of one generalist.
- Token economy by design: AST skeleton extraction and targeted partial reads rather than full-file dumps.
- Knowledge-graph native: queries a dependency graph before falling back to recursive search.
- MIT licensed; installs into any host repository as a git submodule at
.agents/.
🔐 Django-Users-App — installable IAM app for Django
A Django application you install into a project, not a template you copy. It ships a custom user model and the identity machinery around it, and states in a contract exactly what the host must supply.
- JWT and session clients both supported, with token rotation and blacklisting.
- TOTP two-factor with anti-replay, and step-up re-authentication before any sensitive write.
- Application-level encryption for stored personal data, blind-indexed so exact-match lookup still works without decrypting.
- Brute-force lockout per account rather than per IP, through
django-axes. - GDPR erasure that clears every column it claims to, verified by a test that fails without it.
💻 Django-Pro-Template — production Django foundation
An opinionated Django project template that starts from a professional, secure and scalable layout instead of startproject defaults. It is the reference scaffold the other two Django repositories here are built against.
- Environment-based settings with TOML configuration and no secrets in source.
- Dockerised PostgreSQL for local development via Docker Compose.
- Continuous integration on GitHub Actions.
- File-based structured logging wired in from the start.
📈 Tradingview2EXCH — webhook-to-exchange bridge
A Django REST API that receives TradingView alerts and turns them into live orders on Binance.
- Passphrase-gated webhook intake with constant-time comparison and per-scope rate limiting.
- Service Layer wrapping the official Binance SDK. Nothing trades until a dedicated flag says so — never as a side effect of
DEBUG. - Idempotent by
order_id, and precise about it: an order the exchange refused can be retried, one whose outcome is unknown cannot. - Hardened production settings (HSTS, CSP, secure cookies) and a Dockerised deployment.
Open to backend developer roles, and to interesting problems in general.