Data sovereignty
Your prompts, assets and outputs never leave your perimeter. No third-party servers, no data crossing borders, nothing stored outside your control.
@yosoychecoBased in the Dominican Republic
Melvin Checo — Independent Technologist
Self-Hosted AI · Agent Engineering · Technical Solutions
Available for freelance engagements@yosoycheco
- Handle
- @yosoycheco
- Status
- Status available
- Location
- Based in the Dominican Republic
- Focus
- Focus: autonomous systems, self-hosted inference, verified delivery
SEC.02Matrix
Capabilities Matrix
What I ship and production-prove, with honest seniority levels. No inflated claims — every capability maps to real delivered work.
| Capability | Level | Status |
|---|---|---|
AI Agent Engineering Multi-agent orchestration stacks with persistent memory, automatic result verification and autonomous execution over real APIs, databases and tools. | ▮ Expert | Production |
Self-hosted Inference Image and video generation on your own infrastructure — local GPU or VPS — with full data sovereignty and capacity-based pricing. | ▮ Expert | Production |
RAG and Document Automation Retrieval pipelines over private corpora, OCR ingestion, summarization and automated document workflows end to end. | ▯ Advanced | Active |
Backend / API Engineering Next.js, PostgreSQL, Prisma and typed API design with strict TypeScript, automated tests and deployment pipelines. | ▮ Expert | Production |
Linux / Windows Technical Resolution Heterogeneous-environment troubleshooting: drivers, software conflicts, supervised remote sessions and migrations. | ▮ Expert | Active |
Security Hardening Firewalls, 2FA enforcement, fail2ban, audit trails and hardening reviews of servers and applications. | ▯ Advanced | Active |
Level = proven seniority. Status = where the capability is exercised today.
SEC.03Systems
AI Agent Engineering
Complete agent stacks — not demos. Orchestration designed for high task volume, running in parallel, with persistent memory and results that are verified automatically before they count as done.
System characteristics
- Multi-agent orchestration: specialized agents with defined roles, hand-offs and a single supervisor for each workflow
- High-volume parallel execution: many tasks in flight at once, with queues, retries and rate-limit awareness
- Persistent memory: shared knowledge base across sessions — agents remember context, decisions and lessons
- Automatic result verification: outputs are checked against expectations before delivery, with evidence
- Autonomous execution over real systems: agents operate on actual APIs, databases and tooling — not sandboxed mockups
Stack & delivery
What you get, technically
- Built on Next.js 16 + TypeScript strict, orchestrated with the same discipline as any production backend
- PostgreSQL as the persistence layer, Docker for isolation, Linux as the base operating system
- Deployment on your own hardware — local GPU workstation or VPS — data stays under your control
- Oriented to SMEs: workflows that replace repetitive manual operations with verifiable autonomous pipelines
Built for SMBs that need operations to run themselves — scoped, deployed and verified on your infrastructure.
SEC.04Infrastructure
Self-Hosted Inference
Image and video generation with AI running in your infrastructure — local GPU or VPS. You pay for capacity, not per token.
Zero per-token cost
Capacity-based economics: you pay for the hardware, not per request. Unbounded experimentation, predictable infrastructure budget.
No vendor lock-in
Open model weights and standard runtimes. Switch models, swap tools, change vendors — nothing is hostage to a closed API contract.
Compliance-ready
Regulated industries — healthcare, finance, legal — can use generative AI on premise with auditable infrastructure and strict access controls.
Generative AI as an internal capability: private, predictable and compliant.
SEC.05Operations
Technical Resolution
Software problems in heterogeneous environments, solved without guesswork: diagnose first, act second, verify the fix.
Cross-platform troubleshooting
Windows, Linux and macOS. Driver conflicts, update breakage, software incompatibilities — root-caused and fixed, not patched around.
Supervised remote sessions
Remote work done transparently, with your team watching: what changed, why, and what was verified afterwards.
Migrations
Planned moves between operating systems, servers or cloud providers — with rollback paths and post-migration verification.
Documented resolutions
Every case ends with a written diagnosis and the steps taken, so the same problem never costs you twice.
The outcome is a working system and a written record — never just a click and a hope.
SEC.06Portfolio
Case Studies
Selected work, described generically — client names stay confidential. Each case is running or was verified in production.
CASE.01 — Healthcare
Web platform · Payments · Roles
Appointment management platform for a clinic
A clinic needed to stop managing appointments on paper. Built a scheduling platform with real-time slot management, a shared daily agenda and online payments — replacing manual coordination across reception and medical staff.
- Real-time slot availability with conflict prevention
- Shared agenda across reception and practitioners
- Online payment integration with receipt records
- Role-based access for staff and practitioners
CASE.02 — Finance
Next.js · PostgreSQL · E2E-verified
Financial operations platform
An operations-heavy financial product with strict access requirements: granular roles, authentication and production verified end to end — every flow tested against the real deployed system before sign-off.
- Granular role and permission model
- Authentication with session lifecycle management
- End-to-end verification in production, with evidence
- Auditable operation log for every sensitive action
CASE.03 — Automation
Next.js · PostgreSQL · Docker · Linux
Ecosystem of 12 orchestrated AI agents
An autonomous ecosystem where 12 specialized AI agents construct, deploy, QA, security-audit and back up software on their own — orchestrated around real build tooling, with each stage verified before the next begins.
- Agents build features and run their own test suites
- Deployment pipeline run autonomously with rollback
- QA agents execute checks and report evidence
- Security audit and backup agents run on schedule
Generic by design: the work speaks, the clients stay anonymous.
SEC.07Method
How I Work
A fixed, transparent sequence — you always know where the project stands and what comes next.
Scoping call
A direct conversation about the problem, the constraints and the definition of done. Free, no commitment.
Technical proposal
A written plan: architecture, stack, milestones, timeline and verification criteria. You approve before any code.
Build with sprints
Sprint-based delivery with CI gates, automated tests and visible progress at every milestone.
Handoff + optional retainer
Documentation, runbooks and a verified deployment. An optional retainer keeps the system monitored and maintained.
No surprises: scope before build, tests before done, evidence before delivery.
SEC.08Standards
Commitment to Quality
The practices behind every deliverable — because a system is only as good as what happens after it ships.
CI / CD with automated tests
Lint, type checks, unit and integration tests run on every change. Nothing merges untested.
3-2-1 backups, restore-proven
Three copies, two media, one offsite — and restores are actually practiced, not assumed.
Security audits
Hardening reviews, firewalls, enforced 2FA and fail2ban as standard configuration, not an afterthought.
Monitoring & observability
Logs, metrics and alerts that fire when something degrades — before users notice.
Post-deployment verification
End-to-end checks against the live system with evidence captured after every deployment.
SEC.09Direct
Contact
Tell me what you're building — or what's breaking. We'll scope it in a call, no strings attached.
@yosoycheco — replies within 24h on business days