Accelerating Antenna Development

#python#automation#@meta

timeline

A Mesh of Possibilities

As a Principal Engineer focused on Device Verification Testing (DVT), I helped translate signal performance into actionable results using Python, automation frameworks, and systems-level thinking.

A low poly city shows an example mesh network of phased-array antennas connecting across buildings.


Context: What is Terragraph?

Meta’s open-source 60 GHz wireless technology designed to deliver high-speed, low-cost internet in urban settings.

Testing this system is unique. The projects' backbone of leveraging phased-array antennas to establish an outdoor mesh network includes the addition of dealing with non-trivial interruptions from:

  • beamforming
  • signal integrity
  • environment

Terragraph Project Overview & Facebook Connectivity Initiative


Mission: Build Scalable, Repeatable Device Testing

I was tasked with designing and automating end-to-end Device Verification Testing (DVT) for pre-production phased-array antennas.

This entailed developing comprehensive Python automation suites for:

  • RF beam calibration and sweep tests.
  • Spectrum analyzer integration (via PyVISA, SCPI commands).
  • Cross-checks of antenna gain patterns and noise floor response.
  • Data visualizations for heatmaps, antenna patterns, performance analyis.
  • Multi-device orchestration with parallel programming patterns

Documentation That Mattered

I authored internal design specs that defined what "passing units" on the production floor meant, including:

  • Antenna alignment and performance tolerances.
  • Power output ranges and spectral purity.
  • Integration test protocols for mesh behavior and thermal tests.
  • Device bring-up docs and low-level hardware flashing.

These documents became reference standards used by hardware teams across continents.


Data-Driven Debugging

  • Used Python to parse and visualize gigabytes of performance logs, network connectivity, and interference.
  • Automated time-correlated plots comparing RF output vs. control signal logs for deeper root-cause detection.
  • Provided actionable insights to domain experts in timely manner with data visualization and presentations

Global Impact: Taiwan Training & Manufacturing QA

  • Flew to contract manufacturing facilities in Taiwan.
  • Delivered hands-on training to QA and test engineers.
  • Introduced Python-based test workflows to replace error-prone Excel macros.
  • Unified test scripts across remote sites (US, APAC).

Achieved a reduction in test failure rates by over 50% within 3 months.


Reflections

  • RF testing includes art + engineering, especially at 60 GHz, which I love.
  • Working on Terragraph prepared me for my later AR/VR roles (low-level measurement, large-scale automation, cross-team validation).

Automation isn't just faster testing, it's repeatable knowledge transfer at scale.


Want to see how Python can automate complex systems in the real world? Let's connect.