Replacing your smoke test selectors with A.I.

About the Talk:
We can all agree that smoke testing our applications in production is valuable. Tests that fail to complete key interactions is often the first line of defense against regressions and give Engineering teams an edge to minimize or prevent outages that cost real businesses real money.

However, the issue with these tests that rely solely interacting with the frontend of an web application are susceptible to breaking to rigidity in HTML structures, A/B experiments, or even small changes like pop ups blocking elements from being clickable.

This means that every design change potentially means at least one change with your smoke test suite, which can become an issue for Engineering teams trying to juggle the priortity of releasing new features that move the needle, vs technology maintenance on their smoke or acceptance testing suites.

In this short demonstration, Dylan Pierce will show how advancements in Computer Vision & LLMs have opened new doors to replace brittle HTML selectors and rigid line by line instructions with prompts to guide A.I. to complete tasks that provide higher success with less maintenance.

About the Speaker:
Early in my career I cut my teeth building MVPs for clients. I have the technical knowledge to understand the entire stack of web based applications.

I have built a variety of web applications used for niche markets such as social media & marketplace listing aggregation, IoT, Machine Learning implementations for real time judgement, etc.

My focus is on producing viable solutions within the constraints of the timing and resources available. I have worked with startups to release software on a tight schedule.

Including nitty gritty software development, I through mine data to find insights on customer & more recently bad actor behavior. Using this knowledge, I determine the next priority tasks & how to implement them.

From sprint planning, story writing, A/C writing, design, development, writing automated tests, and ultimately deployment. I know the software development cycle intimately and execute it quickly.

I believe in developing LEAN, which favors shipping incremental improvements to accomplish goals over large features that are stuck in WIP (works in progress).

TLDR; I'm a pragmatic technical problem solver and solution builder.

Link to join (Google Meets):