Networking message templates by tech role: SWE, PM, design, data, DevOps

Use a different opening line for each tech function, because each one verifies competence differently. An engineer trusts a GitHub link and a latency number. A designer trusts a portfolio. A PM trusts a shipped outcome, a data scientist trusts a model’s business impact, and a DevOps lead trusts a reliability win. The skeleton of a good networking message stays the same. The proof point you lead with has to change. Below are five role-specific templates and the exact signal each one is built around.

This is part of our referral playbook. If you haven’t nailed the basic structure of a referral ask yet, start with how to ask for a referral on LinkedIn and come back. This post assumes you know the four-part skeleton and focuses on the one part that changes per role: the proof.

The one thing each role screens for

Before the templates, the map. A hiring manager spends a few seconds deciding whether a cold message is worth a reply, and what tips that decision is whether you sound like someone who’s actually done the job. The fastest way to sound like that is to name the proof their function respects. Here’s the signal each role weighs first:

Notice the pattern. Every role wants one concrete thing it can verify, plus one number that maps to the job. The mistake most people make is leading with adjectives (“experienced, passionate, detail-oriented”) instead of the artifact. Adjectives are unfalsifiable, so a busy reader discounts them to zero. A link and a number aren’t.

Software engineering

Engineers read other engineers’ messages looking for evidence you can actually build the thing. The fastest signals: a specific stack with years, a quantified system-level result, and a link to code or a technical write-up they can skim. Drop the buzzwords. “Full-stack ninja” is a tell that you haven’t shipped much.

What lands:

• Specific stack and depth (Go, 5 yrs; React + TypeScript, 4 yrs)
• A quantified, system-level result (cut p99 latency 40%, scaled to 12M MAU)
• One link: a relevant public repo, a merged PR, or an architecture post
• A scope claim a referrer can repeat (“owned the monolith-to-gRPC migration”)

Hi Priya, I’m a senior backend engineer (Go, 5 yrs) applying for the Platform SWE role on your team at [company]. Most recent work: led our monolith-to-gRPC migration across ~200 services, which cut p99 latency 40%. I also open-sourced the retry library we built for it (link). The JD calls out OpenTelemetry, and instrumentation has been my last 18 months. Would you be open to flagging my application, or pointing me to whoever owns the hire?

Product management

PMs are hired on judgment and outcomes, not on tools. So the proof point is an outcome you owned end to end, with a number attached, plus the product area so the reader can place you. A link to a public write-up or talk does double duty: it proves you can communicate, which is half the job.

What lands:

• Product area and segment (B2B SaaS dev tools, mid-market e-commerce)
• An owned outcome with a number (lifted week-1 retention 41% to 55%)
• One framework or methodology you actually used (RICE, Jobs-to-be-Done)
• A link to a public write-up, talk, or LinkedIn post

Hi Devon, I’m a senior PM applying for the Group PM, Growth role at [company]. Six years across two B2B SaaS companies. Most recent work: I led an activation-funnel redesign that lifted week-1 retention from 41% to 55% over two quarters, and I wrote up the framework here (link). The JD centers onboarding metrics ownership, which is exactly where I’ve been. Would a quick refer be possible, or can you point me to the hiring manager?

One wrinkle: PM searches sometimes route through engineering leadership rather than another PM. When you’re messaging an eng manager, shift the vocabulary toward how you work with engineering, not your product wins in isolation.

Hi Sam, applying for the senior PM role on the platform team. I work closely with eng on roadmap and sequencing. My last 18 months were spent partnering with our platform-eng group on API versioning and developer-experience metrics, so I’d come in already speaking your team’s language. Would you be open to flagging my application, or telling me who owns this req?

Design (product designer / UX)

For design roles, the portfolio link isn’t optional. It’s the whole message. A designer who doesn’t lead with work reads as a designer with no work to show. Make the link prominent, point to the most relevant case study by number, and name the surface area you know (dashboards, checkout, onboarding) so the reader can match you to the team.

What lands:

• The portfolio link, up front and impossible to miss
• A named methodology or system (design systems, accessibility-first, motion)
• Cross-functional impact (drove research that changed the roadmap)
• A specific surface (B2B dashboards, e-commerce checkout, AI/ML UX)

Hi Lena, I’m applying for the senior product designer role on the growth team at [company]. Portfolio: (link) — the activation-funnel redesign is case study #2, and it’s the closest match to this role. I’ve spent the last two years on B2B onboarding specifically, partnering tightly with eng to ship the research, not just the mocks. Would you be open to flagging my application to the hiring manager?

Data science / ML

The trap for data folks is leading with model metrics that mean nothing to a business. AUC of 0.91 is invisible to a hiring manager. “Saved $2M in chargeback fraud” is not. Lead with the business impact, back it with the model class and stack, and link a public artifact that proves you can do the work in the open.

What lands:

• The model class you’ve shipped to production (recommender, fraud, LLM fine-tuning)
• Business impact, not just model metrics (lifted CTR 8%, saved $2M in fraud)
• Stack (Python, PyTorch, a vector DB, the MLOps tools you actually used)
• A public artifact: a Kaggle profile, a paper, a model card, or a repo

Hi Omar, I’m a senior ML engineer (PyTorch, 4 yrs in production) applying for the role on your recommendations team. Most recent work: I shipped a two-tower retrieval model that lifted feed CTR 7.2% at [company]. Public artifacts here (Kaggle / paper / repo). The JD calls out cold-start specifically, which was my last six months. Would you be open to a quick refer?

DevOps / SRE / platform engineering

Reliability roles are judged on what you’ve operated and what you prevented. The strongest proof is scale plus a war story: the tooling you run, the size of the fleet, and a public post-mortem or infra-as-code repo that shows how you think under pressure. Cost savings and SLOs are the numbers this function respects.

What lands:

• Specific tooling (Kubernetes, Terraform, ArgoCD, Datadog)
• The scale you’ve operated at (200+ services, 10k+ nodes, a 99.99% SLO)
• Incident credibility (on-call ownership, a post-mortem you authored)
• A number that maps to their pain (drove $1.4M in annual cloud savings)

Hi Rina, senior SRE applying for the role on your platform team at [company]. Six years operating Kubernetes at scale; today I own the autoscaling config and the on-call rotation for ~200 services across four clusters. I wrote a public post-mortem on a 2026 incident (link). The JD calls out cost optimization, and right-sizing drove $1.4M in annual savings for us last year. Open to a refer?

What holds across all five roles

Strip away the role-specific proof and the same four rules carry every message that gets a reply:

1. One specific role.“Refer me to anything” reads as scattered and converts badly. Name the title and the team.
2. One verifiable link. Different artifact per function, but always something the reader can click and check in seconds.
3. One number that maps to the JD.Numbers are the only proof every function trusts, and they’re what the referrer repeats internally on your behalf.
4. Under 90 words.Brevity signals respect for a busy person’s time and makes the ask easy to say yes to.

And the things that sink a message regardless of role: vague flattery about the mission, “looking for new opportunities” with no target, a dump of every tool you’ve ever touched, and skipping the specific role entirely. Pick the two or three signals that map to the posting and cut the rest.

Why this is worth the effort

A role-tuned message is only the opening. What it buys you is a referral, and a referral changes which pile a human ever sees. When someone inside vouches for you, you skip the keyword gate that screens most applicants out, the recruiter gets a warm lead instead of a cold one, and the employee often earns a bonus if you’re hired. Everyone’s incentives point the same way.

Recruiting teams consistently rank referrals as their highest-quality, fastest-moving source of hire. Applicant-tracking vendor Ashby, looking at its own customers’ funnel data, found referred candidates advance through interviews at a markedly higher rate than people who apply cold (see its referrals report). The exact multiple shifts by company and role, so don’t anchor on a single number. The mechanism is the durable part: a role-specific message earns the referral, and the referral earns you a reader. If you’re targeting big-tech or early-stage companies specifically, the contact order changes — we cover both in how to get referred at FAANG and how to get referred at Series-B startups.

Doing this for ten companies without it eating your week

The work behind a good role-specific message is real: find the right person inside the company, figure out which signal their function respects, pull the matching proof from your own history, and track who replied across a dozen searches. That’s the part Rolewynhandles. It surfaces the right contacts by role and seniority at your target companies, drafts the outreach with the signals that fit your function (pulled from your resume and the job description), and keeps a simple board of who replied, who’s pending, and who needs a nudge. You still write the human bits and you still press send. The tool removes the busywork between you and the ask.

Want to see how the referral surface works end to end? Read about built-in referral discovery, or if you just want more ready-to-send scripts, grab the 12 LinkedIn referral message templates.

Frequently asked questions

Why can't I use one networking template for every tech role?

Because each function verifies competence differently. An engineer trusts a GitHub repo and a latency number. A designer trusts a portfolio link. A PM trusts a shipped outcome. A line that proves you to one of them reads as filler to another, so the proof point you lead with has to change even when the skeleton stays the same.

How long should a networking or referral message be?

Under 90 words. Name one specific role, give one verifiable link, attach one number that maps to the job, and make one small ask. If you can't fit the case in that space, the message isn't ready yet. A short, specific ask is far easier to say yes to than a wall of text.

What if I'm an engineer with no public GitHub?

Use a different verifiable artifact. A technical blog post, a conference talk, an architecture write-up, or a specific shipped system with a metric all work. The point isn't GitHub specifically; it's one thing the reader can click and verify in a few seconds. Plenty of strong engineers keep their best work behind a company firewall, so name the system and the impact instead.

Which single link should I include for my role?

Engineers: a relevant GitHub repo, PR, or technical post. Designers: your portfolio, with the most relevant case study called out by number. PMs: a public product write-up, talk, or LinkedIn post. Data and ML: a Kaggle profile, paper, model card, or repo. DevOps and SRE: a public post-mortem, infrastructure-as-code repo, or system write-up. Pick one, not three.

What if my work doesn't have a clean number attached?

Find the closest honest proxy. Scale works (services operated, users served, requests per second). So does scope (owned a migration, led a redesign, set the on-call rotation). Don't invent a percentage. A specific, true claim like 'owned the gRPC migration across 200 services' beats a rounded-up metric a referrer can't defend internally.

Should I open with how much I admire the company?

No. Vague admiration reads as filler and gets skipped. Open with the role and one piece of evidence you can do it. Save genuine specifics about the team's work for when they're real and verifiable, like referencing a system they built or a post they wrote, not the company mission statement.