I run a technical recruiting firm that lives in the weeds of robotics searches. Factory floors, R&D labs, late night standups when the sim refuses to match the real world. I have watched great engineers ignore promising roles because the job post missed the mark. I have also seen lean teams hire fast because they told the story of the work clearly and credibly. This guide is the version I share with clients when we rebuild a posting to reach the candidates they actually want.
Start With the Problem, Not the Org Chart
Top robotics candidates read postings for one reason. They want to know what problem they will own. Replace vague summaries with two or three sentences that make the scope tangible. If the role is about perception for mobile manipulation, say what the robot needs to see, in what environments, and how success is measured. If the role is controls for packaging lines, say which PLC families the plant runs and what cycle-time gains matter this quarter.
Engineers tune out when the first paragraph sounds like a corporate brochure. They tune in when the post reads like an engineering ticket written for adults. A clear problem statement also filters efficiently. People self select in or out based on real signal, not guesswork.
Salary Benchmarks That Set Expectations
Pay transparency wins attention and trust. Ranges do not hurt you. They help candidates evaluate fit and save everyone time. A few current reference points that I share with hiring managers:
- The average base pay for Robotics Engineers in the United States is about $136.6k, with average total compensation near $172k, per Built In’s 2025 data.
- Robotics Engineers mapped under the federal “Engineers, All Other” category show a $117,750 median annual wage for 2024, per O*NET using Bureau of Labor Statistics data.
- Coursera’s 2025 review cites a U.S. average of $98,740 for Controls Engineers with additional pay often significant.
- Robotics Software Engineer median reported at roughly $193k in 2024 based on self reported offers on AIJobs.net.
- Perception Engineer roles can cluster above $200k total compensation at well funded firms, per aggregated profiles on 6figr.
- Electro mechanical and mechatronics technologists and technicians show a U.S. median of about $70,760 as of May 2024, which is useful when you are staffing build, integration, and field service tracks alongside engineering.
Ranges vary by market, sector, and equity. The premium on AI adjacent experience continues to stretch the top end. Reports through 2025 have documented aggressive premiums for strong AI skill sets, which bleeds into robotics software and perception hiring in particular. If your range does not reflect the level of autonomy and scale you expect, the best candidates will infer a mismatch and move on.
One more practical note. Pay transparency is no longer only a preference. States and cities continue to require ranges in postings, and the trend line shows more disclosure, not less. The Minneapolis Fed has a clear summary of policy momentum and trade offs, and New York’s statewide requirement is a good example of where things have landed. When you include a range, say what drives the top of band. Equity and relocation support should be addressed directly.
Define Must Have Skills Versus Nice To Have
Robotics roles mix software, hardware, and systems thinking. Job posts that list every tool someone on the team has ever touched scare off qualified people. Create two short sections. The must have list describes the core of the job. The nice to have list speaks to your stack and your future. Keep both short enough that a candidate can hold them in working memory.
Core robotics software roles
- Must have: Proficiency in modern C++ for production systems, Python for tooling, ROS 2 for messaging and lifecycle basics, CI pipelines, and experience shipping code on real robots.
- Nice to have: Experience with SLAM or visual odometry, learning based perception, motion planning frameworks, Gazebo or Isaac Sim exposure, and skills in optimization or state estimation beyond the basics.
Controls and automation roles
- Must have: Control theory applied in production, PLC or PAC fluency in the platforms you run, fieldbus familiarity, and proven factory acceptance testing and commissioning.
- Nice to have: Safety rated architectures, SIL or PL calculations, experience with vision guided motion, and high speed packaging or palletizing work.
Perception and applied ML roles
- Must have: Camera models, calibration, dataset curation, model training, and deployment on edge accelerators with a feedback loop from field data.
- Nice to have: Multi sensor fusion with LiDAR or radar, 3D detection and tracking, and simulation to support rare scenario generation.
Be careful with time in technology requirements. I often see postings that ask for ten years of ROS 2. ROS 2’s first distributions began appearing in late 2017, with Dashing in 2019 and Foxy in 2020, which makes the request impossible on its face. If you want deep ROS 2 experience, say what deep means. For example, bring up lifecycle nodes, custom message design, QoS tuning for lossy wireless, or rclcpp executor implications.
Spell Out What Success Looks Like in the First 90 Days
Top candidates scan for an honest ramp. They want to see outcomes. Keep it short and measurable. For example. Stand up and benchmark a perception baseline that beats our current model by five points on our primary class. Or, reduce cycle time by eight percent on two target SKUs in the cartoner line by tightening control loops and removing blocking states. These goals invite conversation during interviews and help candidates picture doing the work.
Tell the Truth About Where the Work Happens
Robots exist in the world. Even with heavy simulation and remote tools, field time is common. Be specific about onsite expectations, travel to customer facilities, and after hours test windows when production is not running. Candidates weigh this as seriously as salary. You gain more from clarity than you lose.
Use Language That Welcomes, Not Filters
Language choices in job ads have measurable effects on who applies. The research is not new, but it is relevant in robotics where postings often skew toward hyper masculine tone. The well cited study by Gaucher, Friesen, and Kay shows that gendered wording can reduce women’s interest by signaling they do not belong. Textio’s analysis has long highlighted how phrasing predicts who applies in machine intelligence and related fields. Recent coverage has also called out cliché red flags that signal chaos, like telling candidates they will wear many hats without offering structure.
There is another subtle filter. Several large scale analyses and summaries from LinkedIn and others note that women are more likely to opt out when they do not meet 100 percent of listed criteria, while men are more likely to try at lower match rates. When you write the post, label one short section as preferred or nice to have. Say clearly that candidates rarely check every box. That single sentence invites strong people to raise a hand.
Write Benefits Like an Engineer Would
Benefits matter. State them plainly. Real hardware budgets, simulation licenses, lab resources, conference support, predictable build slots, and a sane on call rhythm. If the job includes equity, define the plan and the vest schedule at least at a high level. If you offer relocation, tell candidates what is covered. If you can sponsor, say so. People weigh these details heavily and appreciate not having to ask.
One Role, One Title, One Stack
Combine responsibilities carefully. It is fine to hire for a product minded engineer who can contribute across the stack. It is not fine to stack three jobs into one posting. If the role expects a person to build embedded firmware in C, write C++ perception pipelines, provision cloud infrastructure, and lead all customer pilots, candidates assume the team is not resourced or organized. Postings like that underperform.
Use Examples To Show the Work Without Revealing Secrets
A few lines of context go a long way. You can share the nature of the tasks without disclosing customer names or IP. For instance, if the engineer will run perception on repetitive pick tasks from totes of small metallic parts, say that. If they will tune force control for citrus handling so fruit is not bruised, say that. People can picture the actual day to day. That alone differentiates your post from the blur of copy pasted templates.
Signal Quality Through Your Interview Plan
The best engineers judge your process as much as you judge them. A fair, efficient process is a positive signal. In the posting, summarize the steps. If your exercise uses realistic data and your interviewers give context, candidates leave with a strong sense of the job and your culture.
Quick anecdote. We supported a perception search for a mobile handling startup. The team shared an anonymized clip of a failure case that stumped them. The interview prompt asked candidates to outline ways to improve the training data and to propose a quick diagnostic experiment. The post mentioned that style of interview up front. Response quality jumped. People who liked problems like that applied. People who did not self selected out early.
Calibrate Requirements To the Seniority You Need
Define scope for each level. Senior roles should own roadmaps and coach others. Staff roles should define patterns and change the way your organization builds robots. Early career roles should learn quickly and contribute within guardrails. I often see posts that say senior but read like principal. That mismatch drags out searches and frustrates candidates.
Use the Right Keywords Without Writing For a Bot
Search still matters. Include the common terms candidates will use, like ROS 2, SLAM, calibration, motion planning, trajectory optimization, PLC, ladder logic, C++, Python, TensorRT, CUDA, OPC UA, FANUC, ABB, UR, Beckhoff. Keep them tied to how you work. A human should be able to read the sentence out loud and nod. If you paste a block of buzzwords at the end, it reads like spam and hurts credibility.
Avoid Common Pitfalls That Push Great People Away
In reviewing hundreds of robotics postings each year, the same patterns knock out solid candidates. These are the ones I fix first.
- Overstuffed requirements. Ten or more must haves will filter the wrong people. Keep must haves to the essentials and move the rest to nice to have.
- Impossible timelines. Claims that a new hire will own production launch in the first month without context read as unrealistic.
- Time in technology traps. Requiring many more years in a technology than the technology has existed is a red flag. The ROS 2 timeline is a classic example.
- Culture clichés. Rock star. Ninja. Like a family. Wear many hats. Recent reporting has called out how these phrases correlate with chaos and unclear expectations. Candidates notice.
- No range. In markets with active transparency rules, leaving out pay ranges limits your reach and may reduce applications. The trend is toward disclosure.
- Three jobs in one. Split the role if you truly need firmware, perception, and DevOps in the same month. Or be explicit that the role is a bridge while you hire two more engineers.
Write The Posting, Then Cut It by One Third
Engineers appreciate brevity that preserves detail. Once you have a solid draft, remove repetition, combine related requirements, and delete any sentence that a candidate could infer from context. If you need to include legal equal opportunity language, keep that, but move boilerplate to the end. The goal is a post that reads quickly and lands precisely.
Put the Team in the Post
Strong people want to know who they will learn from. Name the product or platform, describe your tech stack in a few lines, and link to a short engineering blog post or a conference talk if you have one. If you cannot share names, describe the mix. For example, three senior controls engineers with high speed packaging experience, two perception engineers from agriculture robotics, and a staff software engineer who owns simulation.
Ground the Job in Real Constraints
Robotics teams win by handling constraints honestly. Your job post can do the same. If you run on a tight hardware budget, say how the team compensates with simulation or reuse. If your customers demand validation to specific safety or functional standards, say which ones and why they matter. A simple nod to IEC 61508 or ISO 13849 tells a controls candidate that you understand real world safety work. A perception engineer will want to know how you measure performance and handle dataset drift after deployment.
How I Rewrote a Post That Changed a Search
We once supported a search for a mid level robotics software engineer where the first draft read like a greatest hits list from the entire team. It asked for four years of CUDA, five years of ROS 2, deep learning across five domains, and strong PLC experience. We cut the must haves to C++, Python, ROS 2, experience shipping code to a physical system, and a comfort with troubleshooting in the field. We moved everything else to nice to have and explained the team’s stack. We shared a realistic 90 day plan and clarified hybrid expectations rather than implying fully remote. Applications from qualified people doubled in two weeks. Interviews were stronger because the candidates knew what they were walking into.
A Simple Template You Can Reuse
Use this as a starting point. Replace the content with your own.
Title: Robotics Software Engineer, Mobile Manipulation
Problem: Build and ship perception and planning for mobile manipulation in dynamic warehouse aisles with mixed lighting and reflective packaging.
Impact: Increase successful picks per hour by 20 percent on two pilot lines within six months.
Must have: Modern C++, Python for tooling, ROS 2 with experience shipping to hardware, debugging in the lab and field, proficiency with version control and CI.
Nice to have: Experience with SLAM or VIO, motion planning frameworks, Gazebo or Isaac Sim, camera calibration, edge deployment toolchains.
Location: Hybrid in Pittsburgh, three days onsite for lab access. Travel up to 15 percent for customer pilots.
Compensation: Base range 150k to 185k plus equity. Final offer depends on skills and experience. Relocation support available.
Interview plan: Recruiter screen, technical screen, practical exercise using sanitized logs, final onsite with lab time, decision.
Note on qualifications: Candidates rarely have every preferred skill. If you meet the must haves and are excited about the problem, we encourage you to apply. This single sentence helps counter the application gap noted in LinkedIn’s gender report.
Where Salary Fits Into the Story You Tell
Bring salary into the narrative of the job. Do not treat it as an afterthought in the last line. If your range is competitive for your market and level, candidates will notice. If you plan to stretch for truly exceptional experience in applied ML or perception, say so. That context helps candidates understand where they could land within your band.
If your budget is tight, be candid about the trade. Smaller teams win offers when they sell scope and learning. That is not hand waving. It is a real value proposition for many engineers who want to own a system end to end. Balance the message. Do not promise freedom without support. Say how you mentor and what resources exist.
Final Checks Before You Publish
I end every rewrite with the same quick audit.
- Is the problem statement crisp and specific
- Does the post show the work and the constraints without leaking IP
- Are must haves and nice to haves short and realistic
- Is the 90 day plan measurable
- Is the location policy explicit and honest
- Do the salary range and benefits match the expectations you set
- Did you remove gender coded language and empty clichés
What Great Candidates Remember After Reading Your Post
The best people remember the problem, the impact, and the team. They also remember how your process will respect their time. A clear post sets the tone for a fair interview. It also reduces ghosting because candidates feel treated as partners. You will not hire everyone who applies. You will, however, spend your energy on the right conversations.
I have recruited for robotics roles long enough to know that job descriptions are not paperwork. They are the first artifact of your engineering culture that most candidates ever see. Write them like you write specs. Plain language. Real constraints. Honest tradeoffs. When you do that, the right people show up, and they arrive ready to do the work.
