Having, as my main focus, the objective of entering IME-USP through FUVEST, I developed a study method that I consider effective for the learning of people with a synesthetic mind, like mine: studying by way of research projects.

Here, the term “osmosis” shows its full usefulness. In my preparation for FUVEST, I choose to “hit two targets with a single blow”. Studying biology and chemistry while developing a research proposal in longevity and rejuvenation for VitaLabs seems extremely pleasant to me; in fact, a research proposal that involves machine learning. In other words, beyond the necessary studies in biology and chemistry, the study of mathematics also becomes necessary here, from its foundations to topics such as linear algebra, calculus, probability, statistics, and machine learning.

This, to me, makes studying much more interesting. It is not a matter of studying a discipline only because it will be tested on an exam, even though the exam is, at this moment, my main and immediate objective. It is a matter of making what I study in preparation for FUVEST communicate with questions that will become research, technical projects, essays, notes, code, fichamentos, and better-formulated proposals.

VitaLabs, biology, chemistry, and machine learning

The research proposal for VitaLabs enters this method as studies and biotechnological purpose for part of my preparation. When studying biology and chemistry, I do not want only to memorize cycles, molecules, nomenclatures, or formulas; I want to begin building a minimal language to think, with more rigor, about longevity, rejuvenation, cellular states, risk, cellular identity, and the limits of what can or cannot be inferred from data.

Since I am still at the beginning, this proposal appears more as a study orientation than as a conclusion. The point, for me, is to perceive that subjects that seem distant from FUVEST (machine learning, data science, and biotechnology) depend, before anything else, on very basic foundations: matter, molecule, cell, reaction, graph, variable, hypothesis, evidence, function, matrix, probability.

That is why FUVEST, in this method, is never an obstacle parallel to research. It forces me to study the base seriously. And, if this base is well studied, it can feed by osmosis a future proposal in longevity biotechnology.

ARC Prize and mathematics

In parallel, I see enormous beauty in the ARC Prize Foundation’s proposal for the development of artificial intelligence through Kaggle. Even while needing to maintain rigorous quality, science needs to be a universal space where researchers, from Mumbai to Sao Paulo, from Montevideo to Tokyo, can employ their intellectual capacities so that we may reach, as science, levels that would have been unimaginable 108 years ago.

Here, my studies for FUVEST marry perfectly with studies in mathematics (the same fields mentioned previously) and with fields such as machine learning, LLMs, and AI agents. Some of these topics are not tested by FUVEST, since they are university-level topics; and, even so, the beauty of the method is precisely there: studying for an extremely demanding exam while building, little by little, the foundations that will allow me to think about AI problems with more maturity.

In the case of ARC Prize, this marriage appears in a very concrete way. To understand visual tasks, patterns, grids, transformations, hypotheses, errors, evaluation, and generalization, it is necessary to learn mathematics with proper precision. It is necessary to know how to program. It is necessary to look at a problem and ask: what is changing? What is the rule? What is a pattern? What can I test? What looks like a solution, but may perhaps be only a forced adjustment to a few examples?

How I organize my studies and research

My method of study and research is configured, today, in the following way: first, I use Claude Code to organize local directories where my study and research files will be stored; I use Obsidian for the organization of notes, hypotheses, formulation of questions, and other annotations from my studies. With this, I am able to begin my studies.

My studies concentrate on the following flow: reading books (physical and digital) and papers with real-time annotations, whether in my physical notebook or in my digital notepad; solving questions until reaching 100% accuracy in a set that may vary from 10 to 20 questions (depending on how willing, at the moment, I am to solve questions). The FUVEST exam should not be faced as “I will aim for the cutoff score”, but rather as an exam in which I must seek perfection. Not (only) out of whim, but because of the high quality of the competition, even more so when dealing with the Computer Science program.

The use of flashcards is, yes, used by me, but having, as a hobby, the writing of essays about everything studied so far is, for my learning, something much more powerful than flashcards. Yes, this is where the transformation begins from my performance as a student to a performance oriented toward the production of science.

Essays, regardless of the area, are an excellent form I have found to absorb the content completely while I exercise my abilities in philosophy, sociology, grammar, writing, and the discipline of study in question (in this case, biology, chemistry, or mathematics, but also machine learning, data science, or biotechnology).

What I hope to build with this

Yes, some of the topics mentioned are not tested by FUVEST, since they are university-level topics. Yes, the beauty of this entire study method lies there: not only preparing myself for the entrance exam, but developing my scientific rigor, technical reading skills, improving my abilities as a technology professional, and building seed projects that I hope will be useful to my own future research (and to that of other researchers).

What interests me, at this moment, is to build a type of preparation that leaves traces. Not only hours studied, but better questions. Not only books read, but fichamentos, essays, notes, hypotheses, and small projects that can be resumed later. If I want, in the future, to be a researcher in AI, I need to begin by creating a different relationship with study: less passive, less “school-like” — in the bad sense of the word —, and closer to a continuous practice of investigation.

This is the reason why the blog makes sense to me. It will not be only a place to publish finished texts; it will be a public entrance to the works of my career, even if these works are, at this moment, in their initial phase. I want each article to register a part of the path: what I studied, what I thought, what I still did not understand, which questions emerged, which seed projects began to appear, and how all of this connects with FUVEST, ARC Prize, VitaLabs, and the formation I want to build.