Fifty messages went out in one evening — the same text, pasted into fifty LinkedIn chats, with only the name changed, or not even that. Zero replies. A few messages show as read and unanswered. Then the platform steps in: sending gets slow, an "are you sure this is not spam" prompt appears, and after one recipient reports the message the account is limited for several days. Some students also find that people they later needed — an alumnus, a hiring manager — have already muted or blocked them. The student's own reading of the evening is "nobody helps freshers", which is the belief that keeps the habit alive.
"Hello Sir/Madam, I hope you are doing well. I am a final year B.E.
student passionate about AI/ML with strong knowledge of Python,
Machine Learning, Deep Learning, NLP, Computer Vision and Data
Science. I am looking for opportunities in your esteemed
organisation. Kindly refer me. My resume is attached."
Placement season, or any week when applications are not converting. The advice that produces this is everywhere and sounds sensible: outreach is a numbers game, so send more. Students copy a template from a senior, a coaching class or a social post, and the template is written to fit everyone, which is exactly why it fits nobody. It is most damaging for Tier-2 and Tier-3 students, because their strongest realistic channel is a warm conversation that leads to a referral, and this evening burns that channel across a whole company at once. Linked modules: 6.1 (How Indian companies actually hire), 6.8 (Resume, GitHub, LinkedIn, portfolio).
Four things are wrong, and they compound.
1. Nothing in the message is about the recipient. "Sir/Madam" tells the reader this text was written before they were chosen. A message that could have been sent to anybody carries no evidence of effort, and effort is the only currency a stranger is being offered.
2. The ask is the biggest one available, offered first. "Refer me" asks a person to spend their own standing inside their company on someone they have never spoken to. Many firms tie referrals to the referrer's record, so the request is not small even when the person is kind. A large ask from a stranger is easiest to answer by not answering.
3. The claims cannot be true. Six fields in one sentence, plus "strong knowledge", is a claim no fresher can hold — and the recipient is often the exact person who would interview you. The message damages the later interview it was sent to obtain.
4. Volume triggers the platform, not the human. Identical text sent rapidly is what spam detection is built to find. One report is enough to restrict an account, and the restriction lands on the student, not the template.
The deeper error is a wrong model of what outreach is for. It is not an application; it is the start of a conversation with somebody who has ninety seconds. A message that can be answered in ninety seconds, by a person who already knows something about you, gets answered. Nothing in the template above can be answered at all.
Send five messages, not fifty, and give each one about ten minutes.
Structure — four short parts, under 120 words:
Worked example:
Hi <name> - I read your post about moving churn scoring from a nightly
batch to an API, and the 40 ms budget you had to hold.
I built something much smaller in that shape: a churn model served with
FastAPI, p95 19 ms on one container, and a daily PSI check that caught a
0.31 drift on one feature. Repo: github.com/<me>/churn-repo (runs in
5 minutes from the README).
One question if you have a minute: when labels arrive late, do you
trigger a retrain on drift, or on a rolling metric?
Not asking for a referral - if a fresher role opens I will apply through
the careers page and mention we spoke.
Rules that prevent the account damage. No attachment in a first message; send a link instead, and only send a file when asked. Space the messages out over days. Never send the same first line twice — if you cannot write a different first line for the next person, you have not done the homework for them yet. Keep a log of who, when, what you referenced, and what came back; Career Studio's application tracker is a reasonable place for it.
If the account is already limited: stop sending, wait out the restriction, delete the template, and rebuild the list with five people you can genuinely say something specific about.
Budget outreach in people per week, never in messages per evening — five researched messages is a full week's quota and it outperforms fifty. Write the first line before you decide the person is worth writing to; if that line will not come, move on, because the message would have failed anyway. Build the evidence before the outreach: a repo that passes the clone test and a resume line with a baseline beside the number make a short message land, while no wording rescues an empty link. Ask for the smallest thing a person can give — an answer, an opinion, a pointer — and let referrals be something offered to you. Re-read every message once with one question in mind: could this have been sent to somebody else without changing a word? If yes, it is not ready.
Q: "You wrote to one of our engineers before applying. Why did you pick him, and what did you ask?" Strong answer: name the specific thing you read and why it connected to your own work, state the one question you asked and what you learnt from the reply, and say plainly that you did not ask for a referral. That answer demonstrates research, restraint and the ability to have a technical conversation with a stranger — three things a panel cannot test from a resume. Weak answer: "I messaged many people from the company for referrals." It tells the panel that your process is volume, and it invites the follow-up nobody survives: "what does our team actually build?"