Careers in Software Development
Unemployment among new mathematics graduates keeps reaching new highs. It is increasingly hard to secure a position at a research university. For those mathematicians who specialized in research, software industry presents an attractive alternative.
In 1995 I made the transition from a postdoc at Colorado State University to a software developer position at IBM. This article will detail some of the things I learned during my job search, as well as my impressions after nine months in the industry environment.
Why software?
After years in graduate school and possibly temporary positions, there comes a time to find a more permanent job. For mathematicians specializing in research, the openings are few and far between. Looking for a teaching position is not a good option. There are many new graduates who consider themselves primarily teachers and who have spent years honing their teaching skills. Schools which emphasize teaching will understandably prefer a seasoned teacher to a researcher with a limited teaching background.
The next alternative is to work in industry as a mathematican. Becoming an industrial mathematician is an attractive concept, but those positions are not easy to get. Thus many new graduates find employment writing software.
After spending years in a graduate mathematics program, working as a software developer can look extremely unappealing. Nevertheless, making the transition to software can be very rewarding. First, there is the matter of actually getting a job. The current job market for software developers is good. A candidate with the right skills will be able to pick and choose among employers. There is no indication that this will change in the foreseeable future. Second, there are far more industrial mathematical problems than industrial mathematics positions. With a modicum of luck, a mathematician working in industry will find mathematics to do and problems to solve even if he or she has been hired to do something else. Even disregarding the opportunity to do math, writing software is an interesting and challenging occupation. Like mathematics, software development requires both imagination and painstaking attention to detail. Mathematicians tend to make good software developers.
Preparation
The option of working in industry should be seriously considered early, at least several years before entering the job market. By the time it becomes obvious that search for the academic position has failed, it is too late to start preparing. Desperately replying to the want ads in the Sunday paper, trying to turn a mediocre Fortran programming experience into a job is likely to prove futile. The preparation should ideally cover acquiring both skills and experience.
The basic skill almost universally required is a good knowledge of C. "Good knowledge" means not only knowing how to program, but also how to design software to solve complex problems. A close second is a good knowledge of C++. Many positions require having PC (Windows) or Mac experience, but there are plenty of excellent jobs for people with Unix backgrounds. My background was almost exclusively Unix. I had extensive professional experience in C and a few other programming languages, but no knowledge of C++.
Beyond the programming language and operating systems knowledge, it obviously helps to have as many other skills as possible. With the advent of the Web, perl is a good thing to know. In the Unix world, being able to write X applications will also be a plus. The list could go on and on, but beyond a certain point you are really just trying to show that you are a well-rounded professional, able to learn new skills quickly. It's impossible to know everything.
Gathering experience is almost as important as gathering skills. In hiring fresh graduates, software companies will often restrict themselves to looking at Computer Science graduates. Most positions require experience and writing programs for class assignments does not count. Try to get at least a summer job or two, even a short list of professional projects makes a world of difference.
Types of Employers
I found it helpful to divide potential employers into three categories: contracting agencies, small companies and large companies. Each category has its own set of advantages and disadvantages.
Contracting agencies provide contract labor for other employers. Many times, companies will prefer to meet their staffing needs by employing contract labor instead of hiring additional workers. From the company standpoint, this allows extra flexibility. If the workload decreases, contractors do not have to be laid off like regular workers. Their contracts just do not get renewed.
If you get hired by a contracting agency, it will almost certainly be for a particular contract. When the contract expires, the agency will try to find you a new one. Until that happens, though, you will usually be unemployed. Conversely, sometimes companies switch contracting agencies. In these cases, the contractors will often remain on their jobs, just change employers. Working for a contracting agency has the advantage of flexibility and disadvantages of a less secure position and lower pay. Benefits may also not be as good. Pension plans take some time to vest. If your contract (and you with it) gets passed to a different agency every few years, you may find you never spend enough time with a single agency to participate in a retirement plan.
Working for a small company is probably the best way to gather a lot of experience quickly. Small companies have fewer resources to respond to sudden increases in workload. Some, especially startups, have extremely limited resources to meet even regular demands, so everyone has to work very hard. Stories of 70 and 80 hour weeks are told with some regularity. Job security might also be shaky. A coworker told me of a morning when he came to work to find the doors locked - the company had gone bankrupt overnight. On the brighter side, experience gathered in such an environment might take much longer to acquire elsewhere. And who knows, maybe that startup is the next Microsoft or Netscape, and the stock options will make you rich.
For a mathematician making the transition from academia, working for a large company is probably the best choice. The environment is likely to be calmer, the job security and benefits a bit better than elsewhere. A large company often has different businesses or product lines. Once you are employed, you can get an internal transfer to a position that might suit you better than the one you were originally hired for. There is also a greater chance for finding mathematical problems to solve.
A disadvantage of a large company might be many layers of management, cumbersome bureaucracy and resistance to change. But those things are relative. In today's markets, companies must be pretty nimble to survive. At its bureaucratic worst, a large company will be more flexible than the average university.
The Job Search
"Having worked in academia, what makes you think you'll be happy working in industry?" That was the first question I was asked in the first interview. This point came up over and over again. There is apparently a perception that in academia people do not work very hard and, more importantly, do whatever they like. Thus there is concern that a hire from the academic world will have a hard time concentrating on doing the assigned tasks and instead prefer to sit back and think profound thoughts. Judging by the anecdotal evidence, there may be some basis in fact for this concern. Job applicants must not only have a ready answer to this question, but design their resume and cover letter with this issue in mind.
A separate concern is your legal status in the United States. Most universities are willing to consider nonresident aliens in their searches, and help their candidate of choice obtain permanent residency in this country. As a matter of policy, most industry employers will not consider a noncitizen without a "green card" (i.e., you must obtain Immigration and Naturalization Service authorization before accepting employment). If you don't have a "green card", you will have to search a little harder to find a company willing to do the immigration paperwork. However, this task is by no means impossible.
There have been plenty of books written about looking for a job, networking, researching companies and the like. In February of 1995, when it was obvious that my academic job search was not destined to be a success, I meant to do all of those things. However, I heard stories of people responding to the ads in The Denver Post; and getting multiple offers. Since it was a long time from my last job search, I decided to give it a try simply to gather some experience in writing letters of application and maybe even interviewing. I had decided to stay in Colorado if at all possible, so the ads in the local paper seemed like a good place to start. As it turned out, I never had to go beyond this trial stage.
I sent out three mailings, a total of between 35 and 40 applications. My first mailing contained an adapted academic resume and got very weak response. For the second mailing, I completely rewrote the resume and cover letter. I put great emphasis on my skills and experience, including a separate list of projects. My degree was moved from the prominent place on the first page of my resume to a less prominent place on the second page. The difference in response was startling.
More than ten companies called me back. I interviewed at three places and seriously talked to four companies. In addition to IBM, there was a large defense company, a contracting agency and a medium sized software company.
The defense contractor offered me a chance to keep doing mathematics in my area of expertise. However, the defense contractor was caught in collapse of defense spending and the target date for opening the position kept slipping into future. The medium sized software company was growing really fast and apparently had trouble managing its growth. At one point I was told "The offer is imminent," but it kept hanging fire. The contracting agency needed someone to work on a federal government contract. IBM's offer was far superior in type of work I would do, pay, benefits and prospects. Thus I had no trouble making a decision. Since that time the software company has resumed hiring, but in the meantime laid of some 20% of its work force. If I had taken its offer, I might have already been laid off!
Working in Industry
Industry represents a different set of challenges and rewards from academia. The jobs are likely to be more diverse and less structured. University life ebbs and flows with the school year. In industry, crises and deadlines might happen at any time. In the academic environment, a similar type of work is done throughout career: research, teaching and service, perhaps in varying intensity. In industry, once a project is over, you might be reassigned to do something completely new and unknown.
There is always a lot to learn, and sometimes you have to learn and produce quickly. When I was assigned my first major task at IBM, I had to ask what some of the words meant! I also "inherited" responsibility for quite a bit of existing code that I had to maintain and expand. Being surrounded by people willing to answer my questions helped me get up to speed fast enough to avoid major disasters.
Industry has no tenure, so there are real incentives to try to excel in what you do. Everyone's performance gets evaluated, usually once a year, depending on a company. Contrary to the daily headlines, companies do not lay off employees very easily. If nothing else, severance packages tend to be expensive. However, if your employer does decide to downsize, past performance records have a role in determining who stays and who gets laid off.
An industry employer is more likely to provide adequate resources for doing your job, especially if the business is doing well. In my limited experience, office politics seem to play a less significant role in everyday life. University departments are run at least partially as democracies. In business, your manager will have the last word. You will find, though, that good managers respond positively to initiative by their employees.
Prior to getting an industry job, I shared the widely held view that industry pays far better than academia. After learning more about the pay ranges, it is my observation that the differences are mostly at the lower end of the scale. At a large company like IBM, starting salaries compare favorably with starting salaries in academia. If you are an experienced software developer, the difference between your starting salary and an assistant professor's might be huge. Over time, though, the difference disappears. A full professor at a research university will earn as much or more than a non-managerial employee in industry with similar experience.
Finally, some things remain invariant with respect to the place of the employment: If you do quality work, show intelligent initiative, help your colleagues and keep your customers happy, you will do well and prosper. And you might well have quite a bit of fun in the process.
Nenad Rijavec
IBM Corporation
Science / Engineering courses / colleges
|
