Soumyadeep Dey

So my dev team has gone all in on AI, and it has worked really well so far (surprisingly against the current set narrative) We have really dropped "code" quality bar - but have instead increased "functional" quality of the product by investing more on QA (way more cheaper) We have shipped features almost 40% faster than we used to - and no significant drop in "functional" quality or user reported issues Now we feel we are not moving a fast as we thought it would allow us to - and the reason is code re...

I've been using AI coding tools heavily for the past while, and I've settled into a workflow that works for me. When I let AI go fully agentic on a feature, even with a good spec, I feel disconnected from the codebase. The code is there. It works. But I don't generally get how it works. I just know the result. I don't know what it actually does. And that bothers me. I got burned once. Wrote a short prompt, let AI implement a whole feature, went to test it, and the thing totally diverged from what I wan...

I work for what is functionally a very small company. I was hired at the level right above entry, but even at a low level the responsibilities are high: lead meetings with customers (creating all the diagrams, demos, etc.) be a liaison with customers (communicate with them regarding technical requirements, provide timelines), develop features, and manage deployments. Unfortunately, I've missed artificial deadlines by a very small margin (a day). I attribute this to not focusing on a MINIMAL viable prod...