A Mechanical Engineering Roboticist's Post-PhD Job Search

The future is bleak
Uncertainty's all around
The future is sweet

This is going to be a long post, but I've been meaning to write it for a while now, since I wrapped up my job search in late-December. There doesn't seem to be a lot of posts/summaries for mechanical engineers' job searches out there on the web. I'm not sure why that is. Maybe mechanical engineers just don't like writing stuff. Either way, in this post I'll try to exhaustively detail my job search process the past 3 months or so, since I defended my thesis. In the interest of privacy (and the numerous NDA's I signed), I'm not going to name any of the companies I talked to or interviewed at, and I'm only going to include project details that have already been disclosed online in some form.

My Background

My CV includes the technical details, but to summarize: double-majored in mechanical engineering and computer science at MIT, just finished a PhD in mechanical engineering at Yale, focusing on the optimal design of underactuated robotic hands, with a heavy focus on 3d-printing. Despite my undergraduate degree in computer science, the bulk of my work these past 6 years during my PhD has been mechanically-oriented, so I'd characterize myself as predominantly a mechanical engineer, albeit one with a decent coding background. I'm probably more of a generalist rather than a specialist, unless you need a particular type of 3d-printed robot hand =)

I should acknowledge now that yes, a double-major from MIT probably afforded me a great deal of (at least initial) interest and goodwill that the average job applicant may not have had.

My Process

In short, I literally applied everywhere that sounded interesting for about 2 months straight. I mainly checked LinkedIn, Indeed, and AngelList, as well as tech blogs if they happened to mention a company I wasn't familiar with. I didn't really have any strong geographic, company size/state, or even industry preferences. I was ready to move pretty much anywhere (sorry New Haven), I was open to both startups and established companies, and although it wasn't my first preference, I was open to the idea of a mainly software engineering/coding position.

My networking has always been (and is still) pretty poor, and to compound that issue, I absolutely hate asking for favors, primarily out of pride and stubbornness, so most of my job applications were done pretty traditionally, via the companies' websites or cold-emailing. If you have a choice, this is most definitely the wrong way to apply for jobs. The process is always way easier with an internal reference of some sort. In the end, I did end up reaching out to a few old contacts from my previous internships or conferences, and a couple of friends (bless their hearts) were quick to offer references or introductions, despite my general aloofness to the process.

I should note here that I wasn't in any rush to settle on a job. My grad school life has always been pretty simple and frugal (no car, no relationships, no expensive vices), so I actually had enough directly-accessible savings from my years in grad school to just sit on my ass at my current standard of living for up to 1.5-2 years. I actually strongly considered just pursuing some of my lingering personal projects for a long period of time, whether for a potential startup or just for fun. I'm guessing this is not the norm.

From talking to other folks, it seems that most people start job applications before they defend, while working on their dissertations. I tried that briefly mid-summer but found it to be both frustrating and time-consuming, so I eventually decided to wait until after my defense to pursue the job search in earnest. I defended at the end of August, submitted my final dissertation materials at the end of September, and was basically just applying to jobs and working on personal projects full-time between then and mid-December.

Some Numbers

  • Total job search time: 3-4 months
  • companies applied to: 76+

  • positions applied to: 116+

  • Most # positions applied to at a single company: 13
  • of replies (either yay/nay): 50

  • Average time to initial reply: 29.04 ± 45.69 days
  • Average time between initial phone screen and onsite interview: 37 ± 42.44 days
  • Slowest response: 195 days
  • Fastest response: <1 day (sent app before going to bed, got a reply in the morning)
  • phone (initial) interviews: 41

  • onsite interviews: 10

  • job offers: 6

  • salary range: 85-120k/yr (not including equity/stock where relevant)

The above (as well as the rest of this post) is not meant to be representative of an average job search experience. I'm sure it varies greatly person to person. There certainly were several positions that I applied to on a lark, not expecting to get a response at all, and I'm sure there was also a good mix of jobs for which I was either easily under or over-qualified, resulting in a quick rejection. Difficult to say for sure, though.

Types of Job Opportunities

In my job search, I focused on a few particular types of opportunities: hardware R&D (usually at larger companies), automation and robotics (with a hardware focus), and additive manufacturing. It seemed like these areas were in line with my past working experience and research work. My internships during undergrad primarily dealt with hardware prototyping, some system testing, and actually a significant amount of web development (most of which now is obsolete). Graduate work exposed me to ROS and the basic components for any robotic system (vision, motor control, planning), though not at an in-depth level by any means. My dissertation work involved systems modelling (to propose optimal design strategies), mechanical design (but primarily with 3D-printing), prototyping (to actually build the hands), and final systems integration, testing, and evaluation. I've also been fiddling with RepRap-style 3D-printers on the side since 2013, but that was more notable conversational topic than a substantial skill-set.

All of the postings I applied to could probably be broadly lumped into mechatronics, either the development and testing of the actual hardware as a standalone device, or the integration of several pre-existing technologies to produce some new service. Whether or not it's mostly hype-driven, IoT's definitely huge, not just in terms of wearables or sensors, but in the broader context of increased connectivity and (physical) decentralization. Tangentially, there also seems to be a surge of automation beyond traditional efforts in factories/industry, and not just self-driving cars. I was personally really intrigued by the flurry of attempts to automate foodtech, and likewise was curious how newer entrants to the robotic surgery space would differentiate themselves. For robotics, which I'm defining as automation that handles a significant variety (and not just a controlled set) of scenarios/tasks, the opportunities still lie in the logistics and materials handling space, where you have not only startups trying out new methodologies, but also established automation firms seeking to enhance their existing portfolio with more flexibility and adaptability.

For opportunities related to consumer product development, there was a lot of need for formal DFM/DFA/PLM (which was expected), and really little else. Most of the available work seemed to focus on ramping up or refining mass production, and prerequisites in injection molding and sheet metal bending consistently showed up in the postings I found. Positions for hardware testing and evaluation with a focus on the EE side were also prevalent. I personally found that these typically asked for hands-on experience, and coursework alone was insufficient, so the only available openings would be purely entry-level, which makes sense.

Traditional automation firms had similar requirements, specifically looking for experience with PLC (Programmable Logic Controller), which is certainly very different from what I'm used to with ROS and actually digging into the firmware. A lot of these job postings were really confusing to me, as they had fairly technical descriptions but was mostly just referring to basic industry practices. Again, much like postings for consumer produce development, most of these job opportunities were looking for prior experience specific to certain protocols and their particular industry standards.

Unsurprisingly, opportunities explicitly related to robotics, at both startups and established companies, gave more favorable consideration for graduate level coursework and research, though this was more relevant for software/coding jobs than others. Machine learning, path planning, and kinematics are example topics more likely to be covered in graduate courses or further research. In some cases, especially as ROS becomes more prevalent in industry, there's the opportunity to directly leverage coursework experience. Expertise in machine learning, AI, and machine vision in particular seem like really useful skillsets to have. Whether or not those are just the trendy topics of the moment remain to be seen (imo). Of course, most of this is broad speculation from job descriptions and initial phone screens.

Mechanical engineering opportunities in robotics were a little more frustrating and complicated to find. For design-related jobs, the requirements are similar to those for conventional consumer products. Hardware design iteration is still pretty slow and conservative, for lack of better terms, and to be fair to the hiring companies, experience is pretty key, whether for mass production or designing for the correct expected operating conditions. Many of the posted positions were for characterizing and refining designs for physical, cyclic, and thermal loads. Without getting into more software-oriented tasks like planning, I'd say the most common non-production-related opportunities are in modelling and controls engineering, both of which are more common at larger firms.

There were also quite a few positions related to hardware integration and testing, but most of these were advertised more as technician-level roles, primarily for onsite debugging. Larger automation companies had similar integration roles, (imo) kind of like consultants, where engineers visit clients and propose customized solutions for their needs. The job descriptions for these opportunities were, at least for me, really enticing from a problem-solving perspective. However, while I never got deep enough into the interview process to get a complete view of the responsibilities, it seemed like they're actually much more tedious and systematic, more like routine maintenance and working from a script than really innovating new solutions. Nonetheless, please don't take my word for this.

Startups, especially early stage ones, were the caveats to all of the above. There was typically an appreciation (and often need) for flexibility and someone who was more jack-of-many trades than master of one. That's not to say that this was the case across the board. From what I've seen, as startups grow and make progress, there's definitely a move to specialize more, and experience becomes more important, even if none of the founders had that a priori. In fact, when it came to hardware and robotics, it was pretty common for at least one of the co-founders to have that sort of jack-of-many trades background that you would acquire in grad school. That of course can often beg the question: why would they need/want someone with a near identical skillset as someone already onboard?

Phone Interviews

Initial phone screens were all pretty straightforward, matching the common descriptions found online for all types of jobs. Usually a recruiter or HR representative would spend 30-45 minutes just going over my background and getting a better sense of my background and interest in the company. It typically ends with scheduling a follow-up phone interview with someone in a more technical role. I don't think I ever had one of these end on a negative note, which could be a little misleading, especially since I couldn't that much additional insight into the job beyond the original listing.

The follow-up technical phone screens usually gave me more insight into whether my skillset matched the company's needs, and it was also usually when the companies I interviewed with most aggressively tried to test for competency. I had a couple of interviewers jump directly into a coding challenge or ask fairly specific technical questions related to their work. However, for the roles I applied to, these questions were a little different from run-of-the-mill coding questions I read about for software engineering positions. For anything related to software, the interviewers were often more interested in thought process than any specific coding practices, so we often just discussed answers in terms of pseudocode. Non-software questions were more hit or miss, either I knew them or I didn't. I could answer questions regarding manufacturing, loading conditions, and other design considerations fairly competently (imo), just based on my previous coursework and some intuition, but it was also pretty obvious that I didn't have any formal training/practice in this area. It got to the point that I was just in a constant state of prep, reading whatever I could about mechanical design for review, even though it was basically impossible to predict what I might get asked.

A few companies offered long-term coding or design challenges, over the course of a few days to a week, and I really appreciated these. They gave me the opportunity to review the relevant material more in-depth, and I got a much better flavor of what the company was actually working on day-to-day. That said, I was in the unique position of being unemployed and having no other obligations whatsoever, so these challenges often got my undivided attention.

Onsite Interviews

These were really surprising. I was expecting to get absolutely hammered on the technical end and have my lack of experience get exposed (if it wasn't already), but most of my visits focused on behavioral questions about my potential fit at the company. There were a lot of cookie cutter questions about teamwork and workplace conflict, as well as some discussion of what I was looking for in a career. That said, I'm sure I got into trouble with some (if not many) of my answers to these questions. I was fairly ambivalent about my career and job future throughout my search, and contrary to any recommendation you'd fine online or elsewhere, I was consistently honest about that at each interview. For me, the prospective company's opportunity for growth and investment in me were as important as my skillset's compatibility with the company's goals. Perhaps that was a tad too optimistic on my end, or just a way for me to offload aimless career path elsewhere, but given my lack of urgency in the job search, it was a strategy (if you could call it that) I was comfortable taking, even if it cost me offers.

In many cases, I also spent a good amount of time discussing my thesis work at onsite visits, either in a series of one-on-ones or through presentations (usually just a variation on my thesis defense). Those discussions were a good way to go over my problem-solving approach, some of the side-projects and side-problems I was involved in during the course of my research work, and the different skills I picked up over the years to finish my dissertation. Occasionally, it was also an opportunity for the interviewer and me to commiserate over our collective frustrations with academia.

Discussing thesis work was at times challenging. It's often difficult to determine just how much the audience knows about the subject matter that you've spent 5-7 years researching, or even if they care. Very rarely was there an obvious connection between my research topic and the job in question. Usually, there were certain areas of my work that the interviewer was far more interested in than others, and the conversation often went better once we properly identified that common ground. Over the course of the average onsite visit, I'd often talk to 5+ different people or groups of people, usually from different departments, and so this process would repeat itself several times.

What was the Robotics PhD Good for?

Unless a company is looking to fill a position working on exactly the same thing as your research topic, it's unlikely that a PhD, even in engineering, will provide much of an advantage. For the most part, I was interviewing for entry-level type positions, with the understanding that for most intents and purposes, I wouldn't really know what I was doing once I started. As you can imagine, larger companies are more comfortable with that sort of personnel investment than smaller ones. In fact, all of the offers I got from startups or smallish (<50 employees) companies were directly due to the relevance of my research and/or side projects to their product, and not my skillset in general. Given their limited resources, that certainly makes sense to me.

On the other hand, regardless of what your research topic was, finishing a PhD does mean that you were able to take a problem that no one has addressed and provide a novel solution to push the field further (if only by a little bit), and that's not something very many people can say. That's not to say we've made brilliant contributions to science. I'd venture to say that the most common motivation to leave academia is the lack of sufficient success/notoriety in the field. Regardless, independent research experience is not completely worthless (hooray!).

For me, a robotics PhD with a focus on mechanical engineering gave me a wide breath of experiences across several domains, and I like to think that's the case for most people pursuing graduate degrees in robotics. In my opinion, it may not necessarily make us experts in an area directly applicable to an industry job, but it certainly puts us in a very capable position to quickly learn on the fly when required. If anything, a PhD should teach that pretty much any technical challenge is possible, given sufficient time and a consistent supply of caffeine. Of course, some companies may not value that as much as others, or they simply may not have enough time and resources to take the risk.

The academic publishing cycle and dissertation defense are also pretty relevant alternatives to typical responsibilities like project proposals, integrating feedback (from superiors and customers alike), and fundraising. It's not just about validating the science/design, but also constructing the appropriate narrative optimally for others to understand. Properly communicating the intention can sometimes be more important than its actual execution. As much as we (or just I) joke about graduate students never interacting with other people, most of us should end up doing that pretty extensively, just through the publication process alone. Then again, maybe that's me trying to apply value to the academic circle-jerk.

Certainly, a PhD can also be a cause for concern to certain companies. A few told me explicitly that they thought I would miss the academic aspect of research and get bored with doing work with clients, even though I explicitly and repeatedly stated that research for the sake of research didn't appeal to me anymore. I assume that several companies assumed I would be looking for a higher salary, given my advanced degrees, even though I wouldn't be bringing any additional, directly applicable skillsets. I'm actually kind of curious how many entry-level job postings automatically filter out applicants with higher degrees. I can definitely see cases where the expected value of interviewing a doctoral or master's student is so low that certain companies simply don't bother. For what it's worth, very few commercial (non-research-related) job postings called for graduate-level degrees, though some did count some of the time spent in graduate school as experience.

Making a Decision

When I started my job search, I was heavily leaning towards joining a startup or maybe even trying to pursue some personal projects a little further to see where they might lead. The idea of working on something novel that no one else has addressed is something that still appeals to me a great deal. As it turns out, after talking to a few startups and other early-stage companies more thoroughly, I'm way more risk averse than I thought. There's a significant amount of risk and investment (both monetary and time) to make any serious attempt at a new product/service. There were several startups (even ones I applied to) where I thought the original premise was absolutely ridiculous. That's not to say that I thought they'd fail, but they were ideas that I personally couldn't see myself supporting right out the gate with no validation or traction. My consideration for startups eventually became whether or not the idea would be something I'd pursue even with next to no income, and honestly, there weren't any I found that fit that bill.

With regards to the prospect of starting something myself, interviewing with as many companies as I did gave me a better sense of the gaps in my knowledge base. While I still believe that my graduate work gave me an extensive range of skills, it was quickly obvious that there'd be a significant number of things I'd need to learn from scratch if I wanted to pursue my own consumer product or even a prototype to seek an initial round of funding. In addition, I don't even think I'm attracted to the idea of running and growing a business. I'd much rather continue to work on the actual product/design iteration itself, if I had the choice.

The relative stability and learning opportunities at a bigger company became more enticing as the job search continued. Perhaps ironically, it felt like they were more amenable to hiring me based on potential rather than what I could deliver immediately. Also, while inefficiency and bureaucracy might end up as bad as they (the internet) say, there's also a lot of value in being surrounded by so many seasoned, experienced, industry veterans who most certainly will teach me a thing or two. As it worked out, I finished the interview process with most startups and smaller companies before many of the larger companies even reached out, so I independently decided that the former weren't quite the best fits for me without directly comparing the two.

Final Thoughts

The best jobs aren't posted - Many of my final interviews were at companies that I either initially cold-emailed or got some contact through a referral. In retrospect, this was incredibly frustrating, as I feel like it devalues the applicant's actual technical credentials. Then again, I'm sure it's difficult digging through the deluge of applicants to an open listing, and it's probably both easier and more effective to count on internal referrals instead where possible.

The job search can be long - From the numbers I posted above, you can see that certain companies just took a really long time to respond. That might've been due to a change in the company's HR personnel, an older job posting that just never got updated, a sudden hiring freeze, or a number of other reasons. I'm a little curious if it would've been like this if I had applied during another time period, more in line with common hiring cycles (ie. spring to early-summer). It's certainly frustrating to keep waiting at times, especially if you get offers with tight deadlines.

Many companies don't know what to do with PhDs, and I don't blame them - I mentioned that a few companies told me explicitly that they weren't sure about my motivations and drive for working at said company, and to an extent, I think that's a pretty valid concern. A PhD typically has spent years diving really deep into a single subject matter, so when that person applies to work on something totally different, more than a few eyebrows should be raised. It honestly feels like PhDs either start their own companies/initiatives or just continue to stay in a research-oriented role, usually at a larger company with a significant research division.

Financially, a PhD is a step back - This shouldn't be a surprising comment. Higher education certainly gives you flexibility, but it doesn't trump experience for any particular job. I know plenty of people who made coming out of undergrad than I do now. Unless your research work truly ends up being something groundbreaking, the financial compensation just isn't there.

You'll always answer to someone - A big draw of startups for me was the ability to plot my own path and complete something on my own terms. However, I think that even if you're one of the original founders, you'll still be answering to investors, clients, or consumers. There's no getting around that, and if that's the case, is a larger company with more levels of hierarchy really that much worse?

tldr; Mechanical engineers are not as loved as software engineers in industry. Sorry.