#2025 Summer intern search, only research exp. How cooked am I?
1 messages · Page 1 of 1 (latest)
id link the papers u worked on anyways
obvs dont larp as author
but might help to put
why do u leave off ur math courses
id put em back and stick to comma seperated
ur points are backwards formatted
instead of like "did abd and xyz to reduce jkl by 123%" switch to like "reduced 123% by ..." and bold the metric
since ur pivoting from research to industry
u need to come off as business minded
like they dont want academia oriented folks usually (for most big tech)
like for ur first bullet under each role
use like 0 jargon
recruiter should be able to understand
not only engineers
quick look says this is fine by itself but idk what you're applying for
this is solid for ML stuff but if you're applying to full stack, you don't really have a lot that really proves skills or potential there
and it's very python-centric so bear in mind that it might be an uphill battle if a company is focused on a different tech stack, like java or C++
my field is even more niche than ML LOL
video pipelining, kind of can entail a bit of everything
2024 was terrible for me but i think the market is improving a bit
but yeah my recommendation would be to make a resume geared towards full stack roles
maybe keep experience similar but put in a full stack project or something
if you need extra space, maybe get rid of the TA job, i don't think it really makes a difference
i think for anything research, you really have to sell people on the "why" though
Clean resume.
Here's some pointers.
Some inconsistency in the Coursework section. Write NLP in full if you're not going to abbreviate machine learning to ML. Remove Web Apps and Software Engineering. I would remove all the mathematics courses unless that is relevant to what you're aiming for. It looks like you're aiming for research roles so optionally keep that in. You're listing too many courses though - cut this down a bit.
Second bullet point of first experience - what tech. stack / techniques did you use to enable synthetic data generation? Expand this bullet. Third bullet - I am not familiar with grounding accuracy. I would explain what this is in more layman's terms and also tie how using APIs aided with this (there's a logical leap here). Last bullet needs metric + impact. You fine-tuned an LLM as a simulator and achieved ....
Center your bullets around Google XYZ: Accomplished [X] as measured by [Y], by doing [Z].
First bullet of first experience. Unsure what CIFAR is. Generally gear bullets towards recruiters/HMs. Your second and third bullets need metrics. Try to have a good balance of % and absolute metrics (e.g. impacted over 1,000 developers, saved 400 hours per month of engineer work, developed 4+ API endpoints, etc.) E.g. say you reduced query response times by 40 ms instead of 20%. Or, say you improved query throughput to 400 QPS instead of “increasing by 30%”.
First bullet of third experience. No need to say you researched this paradigm for 8 weeks (implied by the experience dates). This experience would benefit from some introduction to PEFT / FL since these acronyms are used repeatedly. If it was a single bullet (like in the CIFAR case), it's not a big deal. But for someone reading this with limited knowledge of PEFT / FL - a single introductory bullet would be beneficial.
Expand the TA position bullets. E.g. "give students advice on how to divide workload" is not a strong point.
Apply Google XYZ to your project bullets.
I would be more systematic with your bolding. It is a bit arbitrary at the moment. My general advice is either bold metrics or bold tools/frameworks used.
For Skills section, remove Matlab. Remove Software Engineering under concepts.
Overall, solid resume. Main pointers is apply Google XYZ to enhance your bullets. Briefly discuss / clarify esoteric acronyms (like PEFT / FL) to be more geared for layman readers. Be consistent with bolding (either metrics / tooling). Hope this advice was helpful! Feel free to support in bio ☕.
i feel like most the math he listed is highly relevant