Mission Critical Engineer · Security & Infrastructure · Product Engineering
Carlos Ulloque
Engineering for systems where downtime, data loss, and recovery time have real consequences. The work here covers operating Oracle infrastructure under pressure, designing for graceful degradation, and keeping trust and exposure deliberately small.
Focus areas
What the work focuses on
Practical engineering, shared without exposing private systems, clients, or personal data.
Focus
Mission Critical Systems
Continuity and recovery for systems that have to keep running. Designed around failure domains and the path back to service, not just the path that works.
Focus
Security & Infrastructure
Secure defaults, explicit trust boundaries, and a smaller blast radius. Access and privacy decided before they turn into incidents.
Focus
Product Engineering
Software built under real constraints. Controlled complexity, clear ownership, and behavior that stays predictable as it grows.
Recent notes
Field notes from operating real systems
Oracle infrastructure, recovery, and access control — written from incidents and operations, not theory.
Notes
Designing private routes with Zero Trust
What Cloudflare Access actually gates, where the bypass gaps are, and the misconfigurations that quietly leave a route open.
Notes
Cloudflare Access, Java, and JWT validation
Validating Access identity at the origin, with the origin locked down so the gate cannot be bypassed.
Notes
Client-side encrypted systems
Designing so the server holds ciphertext it cannot read — and being honest about the recovery tradeoffs that creates.
Featured projects
Projects built to reduce trust and exposure
Architecture concepts that keep central trust minimal and responsibilities isolated. Experiments stay separate, in labs.
Project
DontTell
Sharing where the key never reaches the server. Central trust stays minimal and stored data is short-lived by design.
Project
Vault
Storage where the provider cannot read the contents. Local ownership of keys, explicit recovery, and no hidden escrow.
Lab
Zero Trust access experiments
Comparing access models for private routes — where each one actually gates traffic, and where it only appears to.
Philosophy
Failure modes come first
Principles from operating systems where degradation and recovery have consequences.
Operating principles
- Decide how it fails before deciding how it scales.
- Keep controls simple enough to operate correctly at 3AM, under pressure.
- Reduce surface area and complexity — smaller blast radius, fewer surprises.
- Favor recovery and predictable behavior over theoretical completeness.