Software, AI and backend that reaches production.
From LLM agents and RAG pipelines to serverless backends on AWS, I design, ship and operate systems that hold up under real data and real traffic. Currently working remotely for a US company, embedded with their team.
Open to roles and projects
AI Backend · Forward Deployed · Remote from El Salvador (GMT-6)
What I do
From model to production. The whole path.
I work across the whole path: architecture, build, deploy and observability, not just the model. These are the four things I deliver.
AI agents & multi-agent systems
Agents that classify, decide and act over your data and tools. Python, LangChain and the OpenAI and Anthropic APIs, with observability, cost controls and fallback paths.
RAG pipelines over private data
Ingestion, embeddings, retrieval and evaluation grounded in your documentation. Vector databases, quality gates and nothing leaking to public models.
Serverless & cloud infrastructure
Backends and pipelines on AWS and Cloudflare Workers (serverless, Terraform, Docker and CI/CD) that scale at low idle cost and survive real traffic.
Full-stack delivery (TypeScript / React)
Dashboards and product surfaces in TypeScript, React and Next.js on top of the systems I build, so your team can actually operate them.
Selected work
In production. Under NDA, told honestly.
Most of my work is under NDA, so here it is anonymized: the domain, what I built, and what it had to survive in production. No client names, no invented numbers.
US insurtech
Full-stack + AIBuilt
Serverless pipelines and React operations dashboards over real policy and transaction volume (AWS Lambda + DynamoDB).
In production
Optimized an operations department: monitored, synced workflows the in-house team owns and maintains after handoff.
B2B consulting
Backend + AIBuilt
Backend and AI integrations for several client companies: multi-agent systems and RAG over private documentation, with evaluation and cost tracking. I ship fast with AI in the loop, but with good practices and review.
In production
Observability, error handling and clean interfaces so non-AI engineers can operate and extend the system.
Web platforms
Full-stackBuilt
Custom web platforms and applications in TypeScript, React and Next.js: dashboards, internal tools and APIs over PostgreSQL and MongoDB.
In production
Shipped and maintained in production, focused on data integrity and clean contracts between services.
I can walk through the architecture and code in detail, under NDA if needed.
Stack
The tools, daily.
What I use day to day to build and run systems in production.
- 01 AI / ML
- Python LangChain RAG Vector DBs OpenAI Anthropic Ray
- 02 Backend & Data
- TypeScript Node.js FastAPI Laravel PostgreSQL MongoDB Redis
- 03 Cloud & Infra
- AWS Cloudflare Workers Terraform Docker Serverless CI / CD
- 04 Frontend
- React Next.js Vue Astro Tailwind
- 05 Bitcoin / L2
- Bitcoin Lightning Liquid RGB
- 06 Workflow
- Git GitHub Claude Code MCP OpenClaw Codex Linux
Specialization
Two uncommon bets. Agentic development and Bitcoin.
Beyond the usual stack, I invest in two areas where few people have real hands-on experience yet: building software with AI agents, and engineering in the Bitcoin ecosystem. Not as a curiosity — these are daily tools, plus training directly with the people building the protocol.
Agent-assisted development
Claude Code and Codex are part of my daily workflow, not an experiment. I build MCP (Model Context Protocol) servers to connect agents to APIs, databases and internal tools, and I run my own personal agents with OpenClaw and Hermes. The result is delivery speed with review and good practices — it’s how much of the work above got shipped.
Bitcoin engineering
In-person technical workshops · June 2026
In-person technical training with developers from Bitcoin Core, Blockstream, Breez, JAN3 and Ark Labs: advanced custody with Miniscript and timelocks, multi-party signing with FROST, Lightning payments (BOLT12, LnURL, Breez SDK), sidechains like Liquid and RGB, and fund recovery and inheritance. I care about Bitcoin as payments and settlement infrastructure, not speculation.
Who I am
I’m Wilmer Salazar.
I’m 21, from Nueva San Salvador. In one to two years I went from Tier-2 technical support to shipping production AI backends for US teams, while finishing a Software Engineering degree. I learn by building things that have to work for real users and real revenue.
Today I work remotely for a US company. Before that: full-stack and AI at an insurtech, backend and integrations for several companies, and custom web platforms. I also explore the Bitcoin ecosystem (Lightning, Liquid, RGB) and run my own agents to stay sharp.
I’m straight about what’s viable and what isn’t. If something won’t work, I’ll say so before time or budget goes into it.
I speak both languages that matter on a team: the technical one and the business one.
- Systems running in production with real data
- Remote with US & LATAM teams
- 20+ courses and certifications (Hugging Face, AWS Educate, Bitcoin Dev, Platzi)
- 15 technical Bitcoin workshops with Core, Blockstream & Lightning devs (2026)
- English C1, embed with international teams
Contact
Open to roles and projects.
If you’re hiring a full-stack / AI engineer, or you have a system that needs to reach production, I’d like to hear about it. The fastest way to reach me is email.