tactical guide: cold-emailing professors as a HS junior (took me 6 weeks to get a yes)

i see this question every week so i'm writing the post i wish i'd read. junior, ISEF finalist, second author on a paper published through a local university lab. here is literally what worked for me. The email itself: - Subject line should name the specific paper you read. NOT "high school student looking for research opportunity". mine was "question about your 2024 paper on X in [journal]". that one got opened. - Two paragraphs max. first paragraph: who you are in one sentence, then the paper you read and one concrete question or extension you thought about. second paragraph: what you can offer (hours per week, summer availability, lab safety training if you have it). attach a one-page CV. - Send tuesday through thursday morning. mondays die in triage and fridays disappear into the weekend. Who to email: - Read 2-3 papers from their group before you write. find the ones where there's a clear next step. PIs notice when you've actually read the paper, genuinely. - Email grad students and postdocs too, not just full profs. 60% of my replies came from postdocs who then looped in their PI. they need help, they're closer to your timeline, and they remember being a HS student trying to break in. - Local R1 universities first. you cannot help a lab you cannot get to. unless you have a real funded summer program lined up (SIMR, SSP, RSI), in-person at the local university beats remote at the famous lab every time. The numbers, for honesty: - 34 emails over 6 weeks - 4 replies, 2 said no, 1 said spring, 1 said yes - The yes was the postdoc, not the PI. she ran the day-to-day, PI signed off after a 15 min zoom. What did not work: - Mass-emailing 50 profs the same template. they can tell. - Saying "interested in your work" without naming the work. zero responses on those. - Asking about "any opportunities". be specific about the project or paper you want to extend. persistence with a bad email is just spam at higher volume. spend the time on the first 10, get the language right, then scale. dm if you want a second pair of eyes on your template, the above is 90% of what i'd tell you anyway.