TL;DR

I want to work on model development, model training, and core research problems at a frontier lab. That is my dream role. For the next 365 days, I’m going deeper into the work I need to understand and showing what I learn along the way.

where I’m starting

I finished my master’s in Applied Data Science at San Jose State in May 2026. Over the last two years, I worked with LLMs, generative AI, and distributed systems through courses and projects. I know how to integrate models and build products around them.

Before that, I worked across apps and systems as a generalist. I won Apple’s Swift Student Challenge and published apps of my own. Moving between different kinds of problems has always felt natural to me. Now I want to use that range to go deeper, not wider.

I’ve been applying for AI engineering roles, but getting noticed for the work I ultimately want to do has been difficult. My goal is to work closer to models and research, and I’m not there yet.

what I need to learn

I need to become comfortable with the mathematics, foundations, papers, model training, distributed training, distillation, and evaluation behind modern AI systems. Reading a term is not enough. I want to understand it, build a working version, test it, and find where my understanding breaks.

The work here will not stay inside one narrow category. It might be a paper reproduction, a training experiment, a voice problem, or a product idea that needs a prototype. The common part is simple: do the work, record the result, and be honest about what I still don’t understand.

I’ll learn from courses, papers, technical blogs, X, lectures, implementations, and conversations. I’ll write about what I understood, tested, built, failed at, or still find confusing.

how I’ll write here

I’ll keep the notes TL;DR-first. AI has made long text easy to generate, and I don’t want to add another wall of words. The result, code, main idea, or thing that confused me comes first. Details stay when they earn their place.

Personal logs may read closer to how I talk—lowercase, direct, and sometimes unfinished. Technical notes will use whatever structure makes the work easier to understand. I don’t want to sound casual on purpose; I just don’t want to polish my own voice out of the writing.

This site will hold the notes, references, implementations, results, failures, and revisions from that work. I hope it remains useful to anyone else trying to move from applied AI toward research at a frontier lab.

In 365 days, I want to be ready for the work I currently consider a dream role.

This is day 1.