Price match and refund automation

At BetterWay Devs I work on automation platforms for US retailers. This case brings together two fronts: a price match and refund bot, and the large scale retail scraping and data processing platforms around it.

The bot has to operate retail sites with real browsers, and do it many times in parallel. That creates three needs: running browser sessions at scale, controlling them from an API, and seeing at any moment how many are running, how long they wait in the queue and how many finish well or get cancelled.

I develop the bot and its control API, built the operations API and dashboard, evolve the Python and AWS Lambda automation platforms and maintain the cloud infrastructure for scraping and data processing. I also modernize legacy code and orchestrate services for high availability, and I speed up delivery with Claude Code, MCP and LLM tooling.

1. Selenium on AWS ECS Fargate, driven by a FastAPI control API

   Fargate runs the browsers without managing servers, and the control API lets the rest of the system launch and watch sessions instead of relying on loose scripts.

2. Infrastructure in Terraform and CI/CD in GitLab

   Describing infrastructure as code means the environment where the bots run can be reviewed, repeated and changed without manual steps.

3. An operations dashboard with the metrics that matter

   Hourly concurrency, queue time and success and cancellation rates show whether the system is saturated or something started to fail, and let the team react with data instead of guesswork.

Running automated browsers at scale is fragile by nature: sessions fail and pages change, which is why success and cancellation rates are first class metrics. On top of that comes keeping high availability while legacy code is modernized and services are orchestrated.

I built an API and an operations dashboard that show hourly concurrency, queue time and success and cancellation rates. The price match and refund bot is still under active development, with its infrastructure in Terraform and its CI/CD in GitLab, and I maintain the cloud infrastructure for large scale retail scraping and data processing.

• Python
• FastAPI
• Selenium
• AWS ECS Fargate
• AWS Lambda
• Terraform
• GitLab CI/CD
• OpenAI API
• Claude Code
• MCP