🤖 AI Product Engineer · Full-Stack · Remote-first

I build end-to-end apps with LLMs that move the business needle

I’m Alex Ariza, a Full-Stack Developer focused on integrating LLMs/RAG into real products: Next.js + Node/FastAPI + managed cloud delivery, with latency, cost, and accuracy metrics from day one.

Download CV

Selected Projects

Selected AI and full-stack projects with demos, implementation details, architecture decisions, and measurable product outcomes.

Featured proof

Production

AI Clinic Operations Copilot — Compliance-First Dental EHR

A production-deployed platform that digitizes Colombian dental health records with append-only folios, certified signatures, FHIR interoperability, and AWS infrastructure.

Cloud deployment

AWS

Certified signatures

100%

RDA / VIDA acceptance

≥ 99%

View case study GitHub

TypeScriptNext.jsFastAPIPostgreSQLAmazon EC2

AI Clinic Operations Copilot — Compliance-First Dental EHR

Recruiter shortcut

Two-minute read: regulatory stakes, append-only clinical records, FHIR RDA pipeline, production AWS deployment, and compliance-by-design engineering.

Internal

SGR/MGA Formulation Engine — Territorial Pre-Investment AI

An internal methodology engine that turns chaotic municipal needs into defendible SGR/MGA pre-investment artifacts — operating inside a real territorial consultancy, with governance that refuses invented data and promised approvals.

Formulation phases

Pipeline steps

SGR / MGA MethodologyAgent SkillsMarkdown + GitOpenCode+2

View case study

Production

AI Internal Knowledge Copilot - Enterprise RAG System

A production-grade RAG assistant that turns internal PDFs into cited, streaming answers.

Illustrative yearly value

9,360 hours/year

Search benchmark

1.8 h/day

TypeScriptNext.jsFastAPIPostgreSQL+4

View case study Demo GitHub

Research

Local RAG Engine — Offline PDF Knowledge Copilot

A local-first knowledge copilot for sensitive PDFs — cited answers from contracts, policies, and manuals without sending document content to cloud APIs.

Illustrative lookup time

Minutes, not hours

Data residency

On-machine

PythonFastAPIChromaDBSQLite FTS5+4

View case study GitHub

Prototype

AI Caloric Estimator — CV + Embedded Prototype

Hardware + CV system that fuses image and weight data to estimate calories with far lower error than manual tracking.

Error Reduction

90%

Accuracy

~10%

ESP32-S3C++FreeRTOSOpenAI API+3

View case study GitHub

Technical Expertise

Full-stack + AI integration: LLM/RAG features, secure APIs, and measurable performance (latency, cost, accuracy) in production-ready apps.

Core Expertise

LLMs, RAG & Prompting

OpenAI APILangChain / LangGraphRAG pipelinesEmbeddings + pgvectorPrompt engineeringRetrieval evaluation

Core Expertise

Backend & APIs

Node.js / NestJSFastAPIPostgreSQL / PrismaMongoDBAuth + RBACQueues & caching basics

Frontend

Next.js (App Router)ReactTypeScriptTailwind CSSDesigning clear UX for AI flows

Cloud & DevOps

Managed cloud platforms (serverless & containers)DockerGitHub ActionsCI/CDAPI gateways & CDN basicsSecrets & access fundamentals

Quality, Metrics & Resilience

Latency & cost trackingLogging/monitoringTesting (unit + integration)Security hygieneTimeouts/retries/circuit breakers

Why Work With Me

Full-stack developer focused on LLM/RAG integration, measured performance, and production readiness.

Shipped AI builds

E2E

Frontend · Backend · Cloud

Full-Stack + AI builder who ships production-ready features

I translate business needs into measurable requirements (latency, cost, accuracy) and build the stack to support them. From frontend UX to backend APIs, vector search, and cloud delivery, I ship features that users can trust.

My sweet spot: LLM/RAG integrations with guardrails, clear observability, and deployment on managed platforms with CI/CD so teams can iterate safely and maximize conversion (low latency, reliable rollouts).

AI/ML Specialist

LLMs, RAG systems, embeddings, and retrieval tuning (chunking, metadata) to keep answers grounded in your data.

Full-Stack Excellence

Next.js + React frontends, Node/NestJS or FastAPI backends, PostgreSQL/Mongo, and managed cloud delivery (e.g., Vercel for frontends, Supabase or a preferred provider for data) so the AI feature ships with the app and performs well for users.

Measured Outcomes

Obsessed with latency p95, token cost, and accuracy. I add logging, tracing, and evals so we know what's working.

💡 My approach: rapid prototyping + production discipline. Speed to value, with security, observability, and sensible costs baked in.

AI Solutions I Build

Production-ready LLM/RAG features shipped end-to-end. Frontend + backend + cloud, with metrics on latency, cost, and accuracy.

Internal Copilots & Semantic Search

RAG copilots that answer from your docs with citations, filters, and relevance tuning.

Use Cases:

•Legal/ops assistants
•Knowledge bases with sources
•Semantic search over internal data

Document Automation

OCR + LLM pipelines to ingest, chunk, extract, and validate structured data from files.

Use Cases:

•Contract review
•Invoice/PO parsing
•Doc summarization with sources

AI-Powered Operations

Classification, routing, and automation using LLMs with guardrails and retries.

Use Cases:

•Ticket triage
•Email intents
•Smart routing + human-in-the-loop

Product AI Features

LLM add-ons for SaaS: draft generation, recommendations, or insight summaries tied to your data.

Use Cases:

•Contextual recommendations
•LLM-driven editors
•Insight cards for dashboards

Quality, Cost & Observability

Measure accuracy, latency (p95), and token spend with logging, tracing, and eval sets.

Use Cases:

•p95 latency budgets
•Token cost dashboards
•Retrieval/LLM evals

Deployments that Ship

Deployment on managed platforms with CI/CD and secrets management, focused on reliable rollouts and conversion-safe releases.

Use Cases:

•API-first delivery
•Staging/prod parity
•Minimal cloud footprint

Don't see what you need? I build custom AI solutions for unique business challenges.

Let's Ship Your Next AI Feature

Ready to integrate LLMs or ship a RAG copilot? I turn ideas into production-ready apps with clear metrics on latency, cost, and accuracy.

Email

arizah2020@gmail.com

Replies within 24h

Location

Colombia · Remote-first (USA/EU overlap)

Open to travel when needed

Availability

Accepting new projects

✓ Ready to start now

Get in touch

Prefer to keep things simple — reach out via email or visit my profiles. I typically reply within 24 hours.

Email: arizah2020@gmail.com LinkedIn GitHub