Daniel Banariba danielbanariba

Hi, I'm Daniel Banariba

💻 Data Engineer & AI Engineer | Tegucigalpa, Honduras

About

Daniel Banariba — Data Engineer and AI Engineer based in Tegucigalpa, Honduras. Currently RPA Developer & Data Engineer at Grupo Farinter.

I build tools that let AI operate production infrastructure, and the data pipelines underneath it.

AI engineering — my differentiator. Six MCP (Model Context Protocol) servers running in production, letting an AI assistant operate orchestration, SQL Server, SMB file shares, deployments and corporate requirements directly. Hardened with a read-only SQL validator, two-step confirmation for production mutations, blast-radius policies with hard stops on destructive operations, and OAuth2. Token consumption optimized ~10–20x. Day to day I work with sub-agent orchestration, Anthropic Agent Skills, slash commands and hooks across Claude Code, Codex CLI and OpenCode — under SDD and strict TDD.

Data engineering. Production ETL pipelines in Dagster over a corporate SQL Server DWH: automated banking and transactional reconciliation via SMB, plus an integration against a bank's H2H mass-payments REST API (OAuth2, secrets in HashiCorp Vault), with file sensors for automatic new-file detection. Cut a monthly demand forecast of ~23,000 series from ~18 hours to ~61 minutes in production — byte-identical, deterministic output, no new dependencies.

Data platform reliability. An autonomous watchdog that detects stuck pipelines by comparing runtimes against baselines learned from 100+ executions, cancels and relaunches them, and alerts over Telegram as a systemd service. Redesigned Dagster failure alerts from send-immediate to freshness-aware, which killed the noise from transient self-recovered failures. I also closed a structural data-quality gap: dozens of asset checks existed but nothing listened to their events, so I built a generic sensor that routes only real errors to Discord, Telegram and email, with cursor-based deduplication.

Power BI as Code. A PBIR/TMDL pipeline with a custom bind-check validator against the official schemas, CI in GitHub Actions, and agentic verification: a watcher that on every push validates, reloads and screenshots the report so an AI can verify what actually rendered.

Backend & systems. hvault — a service written in pure Rust (CLI + daemon) that synchronizes TLS certificates across nodes using HashiCorp Vault (KV-v2), with atomic hot-reload, x509 validation without OpenSSL, and a Leptos/WASM admin UI over OIDC. Also a full-stack competitive pricing dashboard in Python (Reflex): 10 analytical views with SCD2 margin-erosion detection, suggested-price simulation and column-level access control, refactored to multi-country. 1,790 tests, CI/CD.

Earlier. REST APIs with FastAPI and NestJS. VITEK medical-equipment integration over ASTM/HL7 protocols at Analiza Laboratorios Clínicos. E2E test automation with Playwright at GuabaBIT.

📍 Tegucigalpa, Honduras — remote-friendly (LATAM and US time zones)
🌐 Portfolio: https://www.danielbanariba.dev
📧 danielbanariba@protonmail.com
🗣️ Spanish (native), English

#!/usr/bin/python
# -*- coding: utf-8 -*-


class SoftwareEngineer:

    def __init__(self):
        self.name = "Daniel Barrientos Anariba"
        self.role = "Data Engineer & AI Engineer"
        self.location = "Tegucigalpa, Honduras"
        self.language_spoken = ["es_HN", "en_US"]
        self.passions = ["Data Pipelines", "AI Engineering", "Clean Code"]
        self.coffee_level = 100  # Always full

    def say_hi(self):
        print("¡Hola! Thanks for dropping by - hope you find some of my work interesting.")
        print("Code with passion, debug with patience!")

    def refill_coffee(self):
        if self.coffee_level < 50:
            self.coffee_level = 100
            print("☕ Coffee refilled, ready to code again!")
        else:
            print("☕ Still have enough fuel to keep coding...")


# Always caffeinated, always coding
me = SoftwareEngineer()
me.say_hi()

My Skill Set

Languages

Frameworks

Databases

DevOps

QA & Testing

Tools

AI Tools

⚡ Tech Snapshot

mindmap
  root((Skills))
    ((Languages))
      Python
      Java
      TypeScript
      Rust
      C++
      R
    ((Frameworks))
      Spring Boot
      FastAPI
      Reflex
      Astro
      Django
      Express.js
      NestJS
      React
    ((Data & AI))
      Dagster
      dbt
      Polars
      Power BI as Code
      MCP - Model Context Protocol
      Agentic AI
      LLMs
      Anthropic API
      Prompt Engineering
    ((Databases))
      MySQL
      Oracle
      PostgreSQL
      MongoDB
      SQLite
      DynamoDB
    ((DevOps))
      AWS
        S3
        SNS
        DynamoDB
        Lambda
      Azure
      Cloudflare
      Docker
      Terraform
      Vercel
      Kubernetes
    ((QA & Testing))
      Postman
      Selenium
      Jest
      Cypress
      Playwright
      Jira
      Apidog
      Swagger
    ((AI Tools))
      GitHub Copilot
      Claude
      OpenAI
      Gemini
      Perplexity
      Ollama
      DeepSeek
      Cursor
      Qwen
    ((Tools))
      Adobe Suite
        Illustrator
        Photoshop
        Premiere
        After Effects
      Productivity
        Obsidian
        Bitwarden
        Todoist
        PowerToys
      Security
        Tor

Navigation Menu