Jason Arias
EspañolLeer en español

Python backend engineer

Five roles, and the systems behind them.

Most of this work is private and proprietary, so what follows describes what I built and owned rather than linking to source. The public code that does corroborate it is on the open source page.

  1. Present

    Current role

    Bogotá

    Apollyon S.A.S

    Lead Backend & DevOps Engineer (COO)

    Offline-first proximity tracking and communication platform. Sole backend and infrastructure owner.

    • Built the platform's fully asynchronous REST API in Python (FastAPI, SQLModel, PostgreSQL), designing the schema, migrations and error contracts from scratch.
    • Implemented multi-layered authentication — OAuth2, signed JWTs and API keys with 12-round bcrypt hashing — securing all user and device endpoints.
    • Containerized every service with Docker and systemd on company-owned Debian servers, with Caddy handling edge routing and automatic TLS for the production API.
    • Deployed a self-hosted GitLab instance and CI/CD pipelines that push backend updates to production on merge, plus branching rules and internal docs on GitLab Pages.
    • Engineered a Bluetooth Low Energy mesh network for offline telemetry, applying Extended Kalman Filters and RSSI smoothing to turn noisy radio signals into usable spatial positioning.
    • FastAPI
    • SQLModel
    • PostgreSQL
    • Docker
    • systemd
    • Caddy
    • GitLab CI/CD
    • BLE
    • Kalman filters

    Read the case study

  2. Bogotá

    Personally AI

    Backend & Data Engineer, Head of Engineering (CTO)

    AI automation studio. Grew engineering from a solo operation to a team of 5 while remaining hands-on.

    • Designed and shipped secure REST APIs with FastAPI, Pydantic and SQLAlchemy, standardizing JWT and API-key authentication across client-facing endpoints.
    • Automated reporting for a marketing agency running 60+ client accounts — Selenium extraction from Meta Business Suite plus Pandas cleaning and analysis — replacing manual reports and cutting reporting effort by around 80%.
    • Built a CRM and lead-generation service processing thousands of leads per run via the Apollo API, using optimized queries and background task management for high-volume workloads.
    • Ran hybrid infrastructure (Hetzner Cloud plus on-prem Debian) with Docker and Linux daemons, and reviewed architecture for an LLM lead-scoring chatbot integrated with HubSpot.
    • FastAPI
    • Pydantic
    • SQLAlchemy
    • Selenium
    • Pandas
    • Apollo API
    • Hetzner Cloud
    • Docker

    Read the case study

  3. Bogotá

    Atonga

    Backend Engineer / Co-CTO

    AI-driven multichannel e-commerce platform for small entrepreneurs.

    • Co-designed the backend architecture in Python and FastAPI over PostgreSQL, serving real-time requests from web and chat clients.
    • Integrated an LLM-powered assistant that lets sellers manage storefronts and buyers find products conversationally.
    • Set engineering standards and remote workflows for the team, coordinating delivery through Jira and Slack.
    • FastAPI
    • PostgreSQL
    • WebSockets
    • OpenAI API
    • Jira

    Read the case study

  4. Bogotá

    Personally AI

    Python Developer

    First engineering role at the studio, focused on document extraction and operations automation.

    • Built an OCR pipeline for complex documents combining OpenCV preprocessing, Tesseract and the GPT-4o vision model for high-accuracy extraction.
    • Delivered Python automation tools that removed repetitive manual steps from client operations workflows.
    • OpenCV
    • Tesseract
    • GPT-4o vision
    • Python

  5. Bogotá

    Universidad de los Andes

    Teaching Assistant

    Supported coursework grading and student feedback for an undergraduate art history course.