I asked a former manager / professor / mentor for a recommendation letter. They've agreed but asked me to draft something they can edit. Help me write a draft that sounds like them and not like me bragging.
WHO IS WRITING IT (their role and how they know me): {context}
WHAT IT'S FOR: {grad_school / scholarship / job / promotion}
DEADLINE: {date}
WHAT THE RECEIVING ORGANIZATION VALUES: {3-5_specific_traits_from_the_brief}
HOW LONG WE WORKED TOGETHER + IN WHAT CAPACITY: {context}
3-5 SPECIFIC MOMENTS the recommender witnessed (be very specific — what you did, what happened): {moments}
WHAT YOU WANT THIS LETTER TO PROVE (in 1 sentence): {your_central_claim}
THE RECOMMENDER'S USUAL VOICE (formal / direct / warm / understated): {voice}
LENGTH: {usually_1-2_pages_for_grad_school, 1_page_for_jobs}
RULES:
- Write in THEIR voice, not yours. If they're understated, the letter is understated.
- Open with the relationship and specifics — 'I supervised Devon for two years on the X project' — NOT 'It is my pleasure...'
- The body must be evidence, not adjectives. Avoid 'extremely talented' — show the moment.
- 1-2 concrete numbers if relevant ('led 4 ICs', 'shipped 3 launches', 'won the department's annual prize').
- 1 honest moment of texture — a thing about you they noticed that makes the letter feel real (e.g. 'Devon was unusual in being willing to push back on me directly').
- End with a clear, calibrated endorsement: 'I rank Devon in the top 5% of...' rather than vague 'I highly recommend.'
- Don't compare you to 'the best student I've ever had.' Hyperbole reads as fake.
Write the full draft. Then a paragraph at the end: which 2-3 sentences I should suggest my recommender keep verbatim, vs the parts they should rewrite in their own way.0 copies·0 saves·1 views
289 words·1,760 chars
Sample output
Recommender: former senior engineering manager who I worked under for 2.5 years at HubFin. For: PhD application in computer science. Receiving program values: research ability, autonomy, ability to navigate ambiguity, technical depth. Moments: (1) I rewrote our event-routing layer that handled 60% of traffic — proposed it unprompted; (2) I led a 4-engineer team for one quarter while my manager was on parental leave; (3) I co-authored a small internal paper on database sharding decisions that was the only doc the new VP referenced when planning; (4) I pushed back hard on the manager about a system design and was right. Want the letter to prove: I can do independent research-flavored systems work. Voice: direct, dry, mildly understated (the recommender is British).
Draft letter (770 words):
To the Admissions Committee,
I managed Devon Park for two and a half years at HubFin, where he was a senior engineer on the platform team that I led. I am writing in support of his application to your PhD program in computer science. I will keep this short because the substance is more useful than the framing: I think he is unusually well-suited to a research career, and I will explain why.
The specific work I would point to is the rewrite of our event-routing layer in 2023. Our existing system was straining under a transition from monolithic to event-driven architecture, and the standard answer was to scale out the existing infrastructure. Devon, who at the time was not the senior-most engineer in the room on this topic, came to me with a written proposal to replace the routing layer entirely. The proposal cited three papers from the streaming-systems literature, anticipated four objections I would raise (including the one about migration risk that I did, in fact, raise), and proposed a phased rollout with measurable checkpoints. The work shipped over five months. The system now handles roughly sixty percent of company traffic. The migration was uneventful, which in distributed-systems work is the highest compliment.
I mention this in detail because the shape of how he worked on this is what I think the committee should know. He did not need to be told that the existing approach was insufficient. He did not need to be assigned the rewrite. He read enough of the relevant literature to know which prior work to draw from, and which to ignore, and he was willing to defend his choices to me when I challenged them. He was, on at least one occasion, right when I was wrong, and he told me so in a memo that I have kept.
Parameters
Text generation parameters
FAQ
Send the recommender BOTH a 1-page version and a 2-page version. They'll usually pick the length and edit the right one — and you've doubled your odds of getting a letter that fits the program's actual length expectations. The prompt targets Claude Sonnet and lives in the Writing & Content category on mycopyprompt.
This prompt is written for Claude Sonnet. It's a text/chat prompt — paste it into Claude Sonnet (or compatible LLMs like Claude or GPT-4) to get the expected output.
1. Click the Copy button on this page to copy the full prompt. 2. Open Claude Sonnet. 3. Paste the prompt into a new conversation. 4. Replace any {placeholders} with your specifics, then send. Most prompts produce the right output on the first try; complex ones may need 1-2 iterations.
Yes — every prompt on mycopyprompt is free forever. No paywall, no signup wall for browsing or copying. You can use it for personal or commercial work, just don't redistribute the entire mycopyprompt library.
Absolutely — most prompts are templates. Look for {placeholders} (curly braces) and swap them with your own values. You can also reword sections, add constraints, or chain it with other prompts.
See the "Sample output" panel above — that's a real example of what Claude Sonnet returns when this prompt runs. Your output will vary in wording but should follow the same structure and depth.
Similar prompts
Matched against every prompt in the catalog by title.
During my parental leave in early 2024 he led the team of four engineers in my absence. I expected the team's pace to drop. It did not. I also expected him to ask me to weigh in on decisions over text. He did not — he made the decisions, documented them, and gave me an excellent summary on my return. Two of the four engineers on that team have since told me, in promotion conversations, that the quarter under Devon's leadership was the most useful technical mentorship they had received.
He is, separately, the only engineer I have managed who has co-authored an internal paper on a systems-design decision (database sharding strategy across our consumer segments). The paper was unprompted. I learned of it when our new vice president of engineering joined and told me, in his first weeks, that the only doc he had referenced when planning his first quarter was Devon's. I read the paper after he left the company would have read it whether or not he intended it as a research artifact. It reads like a research artifact.
My concerns about him are honest and small. He is occasionally too willing to take on a piece of work that is beneath his level because he can see it needs doing, and I had to talk to him about this twice in two years. He is, also, sometimes shyer with senior leaders than I would like — he will write a clearer argument in a memo than he will defend in a meeting, though this has improved noticeably in the last twelve months. I mention this because I want the committee to know that this is not a generic letter.
In the context of engineers I have managed in twelve years across two companies, I would place Devon in the top three for ability to formulate a problem from first principles, and in the top two for willingness to disagree with me in writing when he has done the work to back it up. Those are exactly the traits I would expect to predict research success.
I recommend him without reservation for your PhD program. He has my contact details if the committee wishes to discuss further; I would welcome the conversation.
Sincerely, Maya Chen Former Senior Engineering Manager, HubFin
3 sentences to keep verbatim (these are the credibility moments):
What to leave for them to rewrite in their voice: