Juan Karlo de GuzmanCall me JK.
An AI Infrastructure Engineer and multidisciplinary systems operator working across Linux, self-hosted LLMs, GPU inference, project delivery, education, legal operations, business development, and executive support.
✓ System understanding improves through controlled contact with reality.
Six professional worlds. One operating system.
Each discipline strengthened a different part of the same capability: understanding systems, coordinating people, handling information carefully, and producing work that survives real operations.
01 / AI Infrastructure
I design and operate Linux-based AI platforms: rootless container services, reverse proxies, isolated networks, model gateways, inference engines, GPU capacity, secret handling, and maintainable production workflows.
02 / Project Delivery
I turn ambiguous initiatives into executable systems through workflow mapping, decision logs, acceptance criteria, stakeholder communication, risk identification, and documentation that survives handoff.
03 / ESL Education
Teaching English since 2019 strengthened my ability to diagnose misunderstandings, adapt explanations, build confidence, and communicate across age, language, and cultural differences.
04 / Paralegal & Legal Operations
Remote paralegal work developed disciplined research, document review, case organization, client communication, deadline awareness, and the habit of separating evidence from assumptions.
05 / Business Development
I research markets and prospects, organize pipelines, qualify opportunities, manage follow-up, and communicate value without disconnecting sales promises from operational capacity.
06 / Executive & Founder Support
I organize information, research, schedules, communication, projects, and follow-through around leaders and remote teams. Running LinguaPath Academy and DelegateOps also gives me a founder-operator view of delegated work.
The systems lab
My strongest learning loop is controlled reverse engineering: deploy a real system, test assumptions, inspect logs, isolate causes, fix the underlying design, and document the result.
Declared services, stable paths, and documented recovery.
Clear boundaries for networks, storage, permissions, and failure.
Understandable by the next operator, not only the original builder.
Architecture follows workload, risk, budget, and operational reality.
Operating method
The same method applies to Linux services, AI platforms, business workflows, client delivery, and unfamiliar operational environments.
Create a controlled working version.
Expose assumptions and weak points safely.
Use logs, documentation, and observed behavior.
Separate symptoms from the actual cause.
Correct the design, not only the visible failure.
Preserve reasoning for operation and handoff.
About JK
My career spans technology, education, legal support, project management, business development, and executive operations. That range is useful when a problem crosses team boundaries, which most important problems eventually do.
I translate business needs into maintainable technical or operational systems without losing the important details.
Running LinguaPath Academy and DelegateOps keeps my thinking grounded in clients, delivery quality, operations, and accountability.
I prefer logs, tests, primary documentation, and observable behavior. Unknowns stay labeled as unknowns until verified.
Written communication, cross-time-zone coordination, documentation, and explicit handoffs are normal parts of how I work.
Professional connection
Connect with me regarding AI infrastructure, DevOps, LLMOps, technical operations, project delivery, executive support, or multidisciplinary remote work.