Hey, I’m Shivam.
I work in GTM engineering around AI security and governance. Before that, I spent years in product marketing, websites, e-commerce, analytics, and growth work. A lot of it was practical stuff. Trying to understand customers, fixing broken journeys, building workflows, getting data into a shape where someone could actually use it.
Over time, I got less interested in only using the tools and more interested in the code, data, permissions, and decisions that power them. That led me through a master’s in Data Science and Machine Learning, and into learning more backend and systems work alongside my day job.
CodeByNight is where I write about that. Forecasting, machine learning, AI systems, backend engineering, Linux, Rust, security, and the place between a model producing an output and someone having to use it in real life.
Some posts explain something practical. Some come from an experiment, a project, or a bug I had to bang my head on for longer than expected. I do find myself asking questions like does this metric deserve trust? What makes a notebook useful outside the notebook? Where deterministic software should do the work instead of an LLM? and How privacy, reliability, or actual user behaviour change the design?
I am still learning, obviously, and I do not write from the position of having every answer. This site is partly a record of what I build and study, but it is also where I try to test my own assumptions.