Need some advice on my future career

[Lip]

Good evening,

I rarely post on forums, but I’m looking for some outside perspective on my current situation.

To keep it short: I am a Master’s student in Europe. My program is focused on theoretical High Energy Physics (covering General Relativity, QFT, the Standard Model, etc.), with some minor courses in computational physics and machine learning. I’ve really enjoyed the program, specifically the particle physics theory and the projects involving data analysis and ML techniques.

However, at the start of my Master’s (right after my Bachelor’s), I faced some health issues. While not life-threatening, they prevented me from attending lectures. I had to self-study while not at my physical best, and as a result, my initial grades were mediocre. My "Intro to Theoretical Physics" grade was "meh," and my General Relativity grade was okay, but not excellent.

Fortunately, my health improved and I was able to attend lectures again. My grades improved significantly; for instance, I recently received a very high grade in QFT, which I’m quite proud of.

As I approach the end of my program, I am considering a PhD. I’m interested in a theory-oriented program—specifically phenomenology rather than formal theory—ideally with a computational aspect. However, I’m worried about my chances. Even if I graduate with a strong final grade, I’m afraid my "sloppy" start, my average Bachelor’s grades, and a 1.5-year gap due to health issues will hurt my applications. I am also a year or two older than my peers.

I would appreciate your thoughts on a few things:

1. Has anyone been in a similar situation? How did you manage the application process?
2. Am I being delusional about a PhD? In US terms, my GPA would be roughly 3.6. Given the competitive nature of HEP, is this realistic?
3. What are the industry alternatives? If a PhD doesn’t work out, I’d be happy working for companies involved in physics or environmental tech. Would looking for a Data Science internship or a similar role be a viable pivot with my background?

Thanks and good day.

 

[Choppy]

For what it's worth, admission committees do tend to take into account poor performance due to illness, particularly when there is evidence that outside of that, the student performs at a much higher level. That said, this is a difficult factor to quantify. What you can do is make sure that it's clear on your application that the health issues impacted your performance and they you have since addressed the issues.
I can't speak to the details of HEP admissions or a data science internship, but it is generally a good idea to have a backup plan.

 

[jedishrfu]

For a PhD you must show a genuine curiosity about your field and a willingness to persevere through courses, and later the research.

My Dad would always remind to be consistent, insistent and persistent. It works well for kids but being a grad student myself it works well.

I think you can make it to grad school. Remember it's part taking the GRE or euro equivalent. Being persistent and insistent to get thru any roadblocks.

For me it was a waiver of the GRE, I’m too old at 73 to subject myself to that test again.

And consistent keeping up with courses, your research, documenting your research and repeating this over and over every day until you graduate with your well earned degree and a few papers under your belt.

I struggle with the consistent part thinking i’ll remember something a week or so later but knowing without a written note I won’t. I also record class lectures and make readable transcripts for self study and exam prep.

 

[symbolipoint]

For me it was a waiver of the GRE, I’m too old at 73 to subject myself to that test again.

That could be interesting for any additional information you can say.

 

[jedishrfu]

Okay, so what happened. I had a student who dropped out the the math PhD from Texas A&M. I don't know why but it sounded like politics or departmental attitude.

Anyway, I learned that a local university had just started a new math PhD program so I told him about it. I noticed that the GRE was no longer a requirement, so I thought, what the heck, I'll apply, realizing that at 73, no one in their right mind is going to let me enter a math PhD. When you're over forty, you're over the hill. No prof would be interested in mentoring someone who will likely produce no papers. It was fun to imagine that I could go.

Then, my second crazy thought, I have a BS in Physics and an MS in CS, so I applied to the CS department as well. I already filled out the application, so that's a no-brainer. As I was getting my materials ready, transcripts, statement of purpose... I noticed the dreaded line GRE required.

So my application was basically DOA. However, inspiration kicked in: "Why not ask if there's a GRE waiver?" The program director said, " Well, yes, there is a waiver, if you can show us a paper of good quality." That threw me for a loop.

I searched through my 45 years of work and came up with two possibilities. The first was an IEEE article on VLSI test system software published in the 1989 IEEE VLSI Test Conference.

Nope, it was a fluff paper, a kind of inventory of software my team and I developed for an in-house test system that ran at 250 MHz for bipolar chip-on-wafer testing.

The second was my CS Master's independent study project on Realtime Programming with two other GE guys. We developed a jet fighter game called Fighter Fox, played on two Atari 800 computers.

I really liked working with the Atari 800 machines but couldn't convince my SO how useful it would be.

The first GE guy says, "I'll handle the joystick and sprites," and after a few weeks, wrote 1 page of joystick reading code and 4 pages of TIE fighter sprites of varying sizes.

The second GE guy said, "I'll write the serial port communication module," and wrote one page of very difficult code because the Atari 800 technical manuals didn't document the APIs very well.

The third GE guy, me, said" Okay, I'll handle the game display" and wrote over 25 pages of assembler code to handle screen repaints, sprite movement, display interrupts, and other arcane functions using a round robin loop triggered by the display interrupt.

There is always an 11th-hour problem, and ours was that the aerial combat sounds interfered with the machine-to-machine communications. We had to disable sound generation. It was due to the Atari 800's architecture, in which the sound chip was also used when you hooked up a modem to generate those classic modem sounds.

We finished our write-ups, merged them, and called the prof in to view our handiwork. He looked at the game for maybe 5 seconds and walked out. We all got As, but it was anticlimactic; we were deprived of seeing the professor play the game.

--- aside

Each Atari had two joystick controls, one stick to control left and right, and up and down. The other stick controlled jet speed and the guns.

We did a display interrupt loop that cycled, refreshing the screen with the other jet's latest position and reading the joystick or the serial port. We made sure we gave enough time to avoid one interrupt interrupting another, which is a death spiral if it continues.

---

Back to the story, I emailed the PhD program director back and said, "Okay, I have a paper." Where do I send it? She said I have to find a professor in the CS department who would be willing to look at it.

I felt like I was thrown under the bus. But you have to persevere, right, so I did a search for the top five CS profs and sent a request to the first professor asking if he'd be interested in reading my paper, so I could get the waiver, and he agreed.

We had a Zoom meeting, and he asked why I wanted to go back to school at my age, and I said I have a curiosity about how things work. He said that's a good answer. and then asked about tuition and said that being self-funded made it easier to get in.

Hand raised, I'm self-funded. He then asked about my patent portfolio and whether I could provide a short, bullet summary of how my patents fit into CS.

Once again, totally depressed, my list has over 150 patents and publications, but no clear grouping. My wife came to the rescue and said, "Take your list and let's see what ChatGPT can do."

Poof, I had the bulleted list sent back to him and sheepishly said I used ChatGPT to generate it. He said no problem, we use it here too. A few days later, he got back to me and said Congratulations, you're now in the PhD program as a part-time student. From there, I discovered that the State of Texas has a Senior Citizen Tuition Exemption program that covers up to 6 credit hours per semester.

Where am I now?

I chose the High Performance Computing path and have just completed my first CS course in 40 years on Advanced Parallel Programming.

All my work experience was of little help; everything was new to me, and the knowledge came at us like water from a firehose. But you must persevere by being persistent, consistent, and insistent: go to class, study, and ask questions until you understand.

---

Since 2005, the world has pivoted away from single-core processors, which couldn't go any faster, to multi-core chips. Parallel programming has become the CS mantra.

I had to unlearn what I knew about threads and timeslicing. Now a thread runs on a core, and you must be concerned with cache coherency and a ton of other considerations.

Basically, I am learning a lot about CPUs and GPUs and how to get the most performance out of the hardware.

---

The key thing I learned was that PhD-level work is really about developing faster or better algorithms, testing them on a variety of hardware, coding them for CPU or GPU cores, and summarizing your results in a paper that you struggle to get accepted by a reputable Computer Science journal or conference.

All this time, I thought CS was about writing cool programs. How dumb could I be?

The next hurdle is a qualifying exam and proof that I can program.

No kidding, the CS department is concerned about graduating someone who can't program themselves out of. paper bag. Your advisor usually signs off on this, or you have to take some lower-level courses to meet the requirement.

I'm on the slow track, one course per semester so far. I may double up in a semester to prep for the qualifying exam, which is offered only in the fall. You need 4 courses over 3 areas of CS. But I'm cautious after my first course. In the fall, it will be an HPC course. From what other grad students have said, these are the two toughest courses in CS.

How's this story?

 

[symbolipoint]

How's this story?

Too amazing. All I can think to do as a response is a Reaction icon for "Wow".