David Bartolomei-Guzmán

Migrating from RunPod to Local Whisper Inference with MLX and a DGX Spark

Sun, 05 Apr 2026 00:00:00 +0000

In my previous article, I described how we use OpenAI’s Whisper model to transcribe radio and TV broadcasts for Monitorea, our media monitoring platform. At the time, we were running inference on RunPod - a serverless GPU platform that lets you deploy ML models without managing hardware. It was the right call to get started quickly. But as we scaled, the economics stopped making sense.

Here’s how we migrated to fully local inference in about a weekend, using MLX on Apple Silicon and a DGX Spark we call Sparky.

Launching Monitorea: AI Agents for Broadcast Media Intelligence

Wed, 01 Oct 2025 00:00:00 +0000

A year ago I published a case study analyzing Share of Voice across Puerto Rico’s AM radio stations using Whisper transcriptions. The article ended with a long list of “future work” — fine-tuning, entity recognition, segment classification, summarization. At the time, those were ideas I wanted to explore. As of this month, most of them are running in production.

Monitorea is now in private beta. Here’s what changed and how we got here.

LLMs Alternatives for Software Development in Regulated Industries

Wed, 09 Oct 2024 00:00:00 +0000

By the time you’re reading this, the landscape of Large Language Models (LLMs) has likely evolved further, but one thing remains constant: their usage in software development continues to ramp up, offering new ways to improve productivity and transform workflows.

Enhancing Software Development with LLMs

If you write software for a living, you’ve probably encountered countless articles sharing how LLMs can enhance your productivity and streamline your development workflows. Tools like Cursor (my personal favorite), GitHub Copilot, and the newly announced OpenAI Canvas are tailored specifically for software development and are great with code generation, error detection, and automating repetitive tasks. If you have access to any of these tools at work, consider yourself lucky. If not, keep reading to explore some alternatives.

Case Study: Leveraging Machine Learning for Spoken Media Analysis – Share of Voice of Puerto Rico’s Political Figures in 2024

Tue, 01 Oct 2024 00:00:00 +0000

This is the first in a series of articles where I share my findings exploring Speech-to-Text (STT) ML models to transcribe and analyze spoken content in news media. In this article, I discuss how STT output can be used for automatic mention detection and tracking metrics such as Share of Voice of political figures in Puerto Rico during the 2024 election season.

The Back Story

Before diving into the details, here’s a brief back story on what sparked my interest in this topic. You can skip directly to the results by scrolling down.

Challenges and Benefits of the Analytics Engineer Role

Mon, 22 Apr 2024 00:00:00 +0000

The Analytics Engineer is a relatively new role that emerged during the “Modern Data Stack” trend of the last 5-10 years and was ”formalized” by dbt Labs. In this article, I discuss some of the learnings, challenges, and benefits that I have experienced while adopting this role in my team.

TLDR: Analytics Engineers merge the analytical skills of a data analyst with the engineering mindset and practices of a software engineer, creating a hybrid profile. This can be very effective in keeping a data team agile, providing insights, and aiding in the decision-making process more effectively than traditional data analysts.

Local Dev with Astro CLI and Remote Databases

Sun, 11 Feb 2024 00:00:00 +0000

When working with Astronomer and Airflow for data processing and ETL tasks, it’s common to encounter scenarios where you need to interact with databases hosted within a private subnet in in your cloud provider. In this blog post, we’ll demonstrate how to set up local development for Astronomer and Airflow while connecting to a Redshift database hosted in a private AWS subnet.

TLDR: We’ll achieve this by adding a new container to the Docker Compose file that contains the Astro CLI configuration and configure the Airflow Connection to use that container to reach to your remote DB.

Managing Python versions and environments in macOS

Wed, 01 Dec 2021 00:00:00 +0000

I have been using Python for a while now, and managing dependencies and versions always has been a pain. macOS still comes with Python2.7, but most dev works have moved to Python3.

Graphic representation of a common Python setup.

This time around, I saw myself using Python more and more and decided to stop kicking the can down the road and “properly” set up Python (at least for my use cases). This is a quick setup summary on how I set up and manage my Python versions and environments using pyenv and pyenv-virtualenv. These tools are also compatible with various Linux distributions such as Debian, Ubuntu, Mint.

How to easily migrate your Brew packages in macOS

Sun, 06 Dec 2020 00:00:00 +0000

Recently I decided to do a clean install on my personal Mac at the same time and start optimizing my dev setup and overall productivity. The first step towards this goal was to automate installing all the packages and apps I use daily. I have been using Homebrew to manage my macOS dependencies for a few years now and recently I discovered the awesomeness of Brewfiles to export and move dependencies to your new Mac and/or to track it in version control.

From Software to Data Engineering

Mon, 26 Oct 2020 00:00:00 +0000

I started to work as a data engineer when me and my team were tasked to automate some data sets and reports for the businesses/brands our team was supporting. The challenge was significant: these sites lacked any established data pipelines, relying solely on basic event capture. What was initially a three-month project ignited my passion for data and analytics. In this post, I’ll share my experiences and key learnings from the past year.

David Bartolomei-Guzmán

Mon, 01 Jan 0001 00:00:00 +0000