A hackathon win used to prove you could build under pressure. AI agents made the execution cheap. What still counts as proof then?
What it costs to enter research right now, who is quietly paying it, and which parts of the work still have to be yours.
The math that determines your vLLM deployment's true concurrency before an OOM teaches you the hard way.
Notes to myself, on doing things badly, for no one in particular.
The quintessential blueprint for becoming a thoughtful and impactful AI researcher.
Building an Interactive Web Demo showcasing Touchless Interactions using the MediaPipe Machine Learning Library.