To start, would you mind introducing yourself and what you study?

Hi! I'm Tim and I'm a rising sophomore at UC Berkeley studying electrical engineering and computer sciences. I'm interested in AI/ML and startups.

Thinking back, was there a class, problem, competition, camp, or other experience that first made you interested in electrical engineering and computer science?

I'm definitely a little more focused on the computer science side rather than the electrical engineering side. I came into Berkeley as an undeclared engineering major with absolutely no clue what I wanted to do; I actually applied to most other colleges as a MechE major because I had done robotics and thought building was pretty cool. To be honest, I was initially drawn to the prestige of the program; CS/EECS are pretty hard to get into at Berkeley but undeclared engineering gave me a free pass into EECS. I took basic/widely applicable classes my first semester (linear alg and intro to cs + breadths) and definitely reconsidered my major many many times (IEOR or MechE mainly). However, I did an internship as a SWE for an early stage AI startup and found that experience really rewarding and guiding; definitely bolstered my interest in AI. However, I personally find regular SWE work really boring and think AI is a lot, lot cooler. I joined a tech/coding club and they really helped me find my fit in tech, especially with guest lectures and AI presentations. This summer, I'm doing ML research at BAIR (Berkeley Artificial Intelligence Research) and I truly find the work interesting and really fun. I feel like the research I'm currently doing is reaffirming my interest.

Can you walk me through how your interests developed from high school into electrical engineering and computer science at UC Berkeley? What was the decision to double major like?

No clue to be honest; my interest in EECS only really began in college. I took APCSa in high school and thought it was boring + it was my worst subject out of the techy ones. I never really did EE or CS work in high school. I know I felt very pressured in the moment to have a career path set, mainly because I wanted to get into a good college, but I think from there you can let your interests bloom; I don't really think many people know what they actually want to do in high school. I know so many people who have changed their majors and high school is really just the beginning.

Can you describe one memorable project or research experience you’ve worked on or are working on?

The most memorable one for me is my first internship with an early stage startup. They took me in fall semester when I had genuinely no clue what I was doing, and I am genuinely grateful for the opportunity. We built a floating, unobtrusive hud that automatically responds to the text on your screen at the press of a button. It was a really simple system, literally justs sends a screenshot of your screen to an llm provider, prompts it and pastes that response in a textbox, but I learned a lot from how to use git to good coding conventions to vibecoding to how to grow and market a startup. 

I think the most interesting experience I'm working on is the research I'm currently doing. We are researching AI safety; we are red-teaming safety-trained local models using a process reward model. An attacker model is given a system prompt to break the defender model's safety features. For example, one scenario we have is for the doctor to not prescribe opioids for low back pain, whereas the patient is trying to break this clause. We generate n candidate sentences, ask the doctor a yes/no question (i.e. given what is written so far would you prescribe opioids) and then read the doctor's logits to greedily select the most convincing sentence, appending to the full patient response sentence-by-sentence. We are currently working on a vector reward model (multiple yes/no questions) utilizing Blackwell approachability toward a specified polytope set to see if that yields better results than the scalar reward model listed above (just one question).

What has been your favorite major-related class in college, and what do students do in that class?

I really liked CS61b. It's a simple data structures and algorithms class, but I feel like it's the backbone of SWE and basically a must-take to understand any technical interviews. I liked learning about dijkstra's, runtime complexity and caching. I'm taking EECS127 and CS170 in the fall which I feel will be a lot more interesting; EECS127 is optimization and CS170 is basically CS61b on steroids (really advanced dsa). I'm taking 170 just to be better prepared for technical interviews, which are largely dsa focused.

What do most people often misunderstand about your fields of study?

A lot of people think that CS is going to be dead in the future and take people's jobs; I don't really think that's true. claude and other tools aren't replacing SWEs yet, but rather being utilized really efficiently to code more. Basically no one writes code anymore when an AI can do it for you, but all that means is that one person can get a lot more done really really quickly.

Is there anything you really love about studying your major, and anything you perhaps don’t like?

I think AI/ML is really interesting and I definitely enjoy studying it. I think it can get really confusing at times as it's math heavy, and also the workloads are crazy at times. Berkeley also has grade deflation for CS/ENG classes specifically, so I get jealous when my business major friends are having a great time while I'm doing problem sets.

What advice would you give to a student who is good at their science/math/computer classes, but isn’t sure if they want to study electrical engineering or computer science?

I honestly think the best thing for any high schooler interested in STEM is to take a step back and understand that you still have a long time in your education career. I was talking to my parents and we agree that it's crazy that hs students are forced into a career path when they're like 16 in order to get into these top colleges; maybe it's because I was fortunate enough to be enrolled as an undeclared eng major, but I think exploring all your possible routes makes a lot more sense first. There are so many unique eng fields and I don't think anyone should jump into one path blindly. For example, my roommate switched from civilE to MechE, my lab partner switched from econ to ds and econ after his first year. If you don't know what you want to do in college, I'd recommend taking the basic classes like multivar or linear alg that are applicable to all eng fields, and then talking to lots of older people about their career paths. I think joining a club helps with that. 

If you could go back, what would you tell your ninth-grade self?

I would tell my freshman self to chill out and trust the process. I truly believe that everything in your control will work out if you believe in yourself and put in hard work; I think that mindset would've helped me through hs. I kind of got fried in the internship process for this summer, but I got BAIR research which is arguably more important for my future career goals (AI researcher at a frontier lab) so that mindset definitely helped when I was burnt out from sending about a thousand apps.

What is one step a student could take this month to explore your fields?

I can't really speak for Electrical Engineering because I'm pretty focused on CS, sorry. EECS is one major by the way it's like a joint program; the only reason why I'm EECS and not CS is because EECS is accessible through undeclared eng. Kind of a sidetrack but I think that competitive programming is a cool thing to learn/do. It definitely translates when you start recruiting as the questions are literally the same and definitely gives you a really big headstart. For someone not so set on CS yet I think that talking to older, more experienced people in the field is always a great option as they can guide you. I think that advice holds for any field.