why is it too difficult to get a job in the so called CFD?

I have been looking for a job in this area for a long time and so far no one even responds when I send them a resume. I suppose that there are so many people working in the CFD area that there are not enough available jobs in industry or in universities. If anyone has deeper insight into the CFD job market please continue the dialogue.

(1). CFD was largely associated with aerospace industries, especially the defense industries in the past. (2). With the cut-back in defense spending, merging of aerospace companies, it is hard to find any job in aerospace industries, not just CFD job. (3). Several hundreds thousand of aerospace engineers and workers have been displaced in the past ten years and the future is still uncertain. (4). But new areas such as Internet and computer animation have opened up new jobs in the last couple of years. Every company thinks that they must have a web page in order to be competitive in the marketplace. ( No one think that they must have CFD in order to survive.) Even for the commercial CFD code developers, the most demanding job is in the sales and marketing. (5). It is a good idea to learn some Internet skills. Currently, the main market for CFD is largely gone. The global political environment has changed since early 90's. The rapid growth in Asia is also gone. Only the person who has real skill and interest in CFD can survive in the future. (6). The future in CFD is not bright right. When there is no problem to solve, the code is useless. In many areas, CFD is not used to design new products but to support service problems ( it takes money to solve problem, and it also takes money to answer questions here . There is no free lunch even at CFD forum.)

This is certainly true if you look at the CFD job-market in a longer perspective (several years back). However, I'm not sure if it is this NASA/governmental-aerospace cutback that is the important factor now. Half a year ago the market for CFD engineers was better, but lately big players in the US (automotive, aerospace, Fluent, adapco, CFDRC etc). seem to be hiring with more caution.

The scene is different in Europe -AEA, Computational Dynamics and others seem to be hiring a lot now. I also know that several of the larger aerospace industries in Europe have difficulties to find CFD engineers.

I think that it is just a sign that the US is ahead of Europe in this economic cycle - US has started to flatten out but Europe is still rising.

(1). A company hires an engineer because the qualification matches their job description. The job description is written by the department manager, so he is not going to say that his department is lack of the CFD technology. He normally would say his department has several CFD codes, such as ....etc, including commercial CFD codes. In some cases, he will tell you that he knows how to save the money for the company by using a commercial CFD code with unlimited licenses ( naturally through his effort of cost reduction study...). So, the only reason why he needs to hire another engineer is because the old engineer in his department has just retired, and his department has been working very hard to meet the time schedule, and he is trying to improve the productivity to meet the customer's demand, to meet the company's goal, etc... So, he claimed that all of these goal will be met if he can get an additional engineer. Therefore, if you say that you are a hard working fellow, and always knows how to write a "small" computer programs to get the results on time, then you have something in common. See, a manager is measured by the resources he has. So, you can run more CFD cases for him, he will be able to get more results and thus more budget. And If you say you are interested in CFD research and code development, he is going to have hard time to sell the idea to his boss, because you do research when you have extra money to spend, unless he is a research type. I think, if you read through the job listings you can easily get a conculsion that they are looking for engineers to work and make a profit for them. So, the important thing to say is that you understand their need and you can perform the job well in many ways. The computer related skills are always very important. That is the reason why it is very important to write your own CFD code. You will be hired but it is likely that they are impressed by your computer skill, not because of your code. ( They already have too many codes on their systems. or they want you to run a particular code.) (2). In old days, the government has a huge defense budget to build new aircrafts, fighters, missiles,etc. They require the best quality.( the cost is not the major concern) The goal of CFD was to obtain the best results using the best available technology and resources. Unfortunately, those days are long gone.

John, Thanks for enlightening me on the CFD job market. I understand what you are saying, unfortunately, I am one of those Ph.D. students in which writing my own code from scratch and performing a study on a CFD problem is the requirement for graduation. I have the knownledge in CFD as far as writing code, theoretical formulation of the problem, etc., but I do not have the skills in order to get a job (i.e., knowing how to use a popular commerical CFD package and calling myself an expert). It seems to me that universities (at least not mine) are not really preparing there graduates for industry. I might seem pessimistic at first, but I think in the long run (over many years) the education will be well worth the time and energy invested. Thanks for keeping the dialogue going.

CFD should be considered a tool for solving engineering problems. A knowledge of the application to which CFD is to be applied would certainly be advantageous in obtaining a job. A rather simplistic analogy would be the use of MS Word. When applying for a job it is of little use in saying that your are profficient in the use of MS Word if you can only write Chinese poetry but the job requires you to create technical documents for the US market.

This is not true of course if you are applying for commercial CFD development post. I imagine though that there are at least 1000x more people using commercial CFD codes in industy than there are writing the codes themselves.

Robin.

The difficulty with finding a "CFD job" is one that I encountered when I finished my Ph.D. at the end of 1994. I cannot comment on the CFD job market in general, but I think that some of the difficulties associated with finding a CFD job are the same as would be encountered with any job search.

When I completed my Ph.D. I had no industrial experience, except for some summer co-op jobs. For jobs in many fields, including CFD jobs, candidates with previous industrial experience are strongly preferred.

The other important issue, which I cannot emphasize strongly enough, concerns job search techniques in general. Like Chuck, I have had the experience of replying to a job ad that sounded like it was written with me in mind and not even receiving a letter to acknowledge the receipt of my resume. One recent example for me occurred last summer. I replied to one of the job ads posted on the CFD Jobs Database. I had the experience sought, I had been working in the same industry, and I was already within driving distance of the company (no travel costs to interview or relocation costs to move for the job). I never received a letter or a call from the company, and the job ad remained on the CFD Jobs Database for several months after I applied for it!

The point that I am trying to make is that replying to job ads, whether posted on the Internet or listed in the newspaper, is very unlikely to yield success. I'm not saying that you should not bother to respond to job ads, but I have heard estimates of the success rate as being only 2-5%. I have also found this to be true myself. When you respond to a job ad, your resume is probably going to be one of dozens or hundreds that winds up on the desk of someone in a Human Resources department who has no knowledge of CFD and no direct connection to the advertised position. You need to reach the hiring manager or someone in the company who can reach the hiring manager.

Much better results can be achieved through networking. All of the interview and job opportunities that I have had have been generated through networking, talking to people who I knew or even people who knew people that I knew.

I apologize for the long posting. The point I am trying to make is that finding a CFD job is at least as much about finding a job as it is about CFD. Please contact me if I can be of any help with job search techniques. I have a lot of personal experience and information on the subject, and I recently gave a presentation about job search techniques at a technical conference. Good luck to all who are searching for a CFD job.

Elliot

(1). In old days, the engineer using the CFD code is the only person creates the code and modifies the code. There is no other alternative. (2).In 90's, companies have the choice of using a commercial CFD code without developing their own in-house code.(whether this is the right approach or not is not settled yet) (3). A company with tight budget can now access the CFD field through the use of a commercial code. And with the CFD vendors anxious to increase their market shares, a couple of weeks intensive training is all that is required to get an engineer start running the code. (4). It takes several years of training to create a PhD in CFD field, but the job in industries simply disappear in two weeks. (5). This is acceptable in a compnay which is more interested in the cost reduction than in the CFD technology. So, the CFD job is now replaced by the CFD code and engineers with two weeks of training. (6). I am not ready to comment on this new trend because it is not my responsibility to change the course of a company , any company. (7). All I can offer is my experience. Back in early 80's, I had an opportunity to use a large finite-element CFD code with arbitrary geometry capability. I was able to get the geometry and the mesh, but not the converged solution. In early 90's again, I had another opportunity to use another finite-element CFD code , this time , the same thing happened again, the solution could not converge in several months. (8). I am running a code ( a well-known code with source listing) with several layers of improvement in it. At the same time, I have discovered several layers of bugs in it. These were included because the engineers worked on it were not professional CFD engineer. (9). Along the commercial code line, the same thing also happened. In most cases, it is related to the geometry and the mesh because they can not change the solver part. And the engineers were always forced to use the non-converged solutions. They also tried to fine tune the mesh in such a way that the results will make sense to them. All these were acceptable because it was nothing but analysis or simulation which could not have any bad impact on anyone. (10). My prediction is that in the next century, the PC will be very popular ( it is already very affordable nowadays), and the real CFD engineer will be able to take advantage of the situations to write good CFD codes and have a big impact on industries and nations as well. Based on my observation, companies without real CFD engineers all have troubles already. (11). The survival of a company in the next century depends solely on the ability to keep an active force of real CFD engineers. If I am running a code with two additional generations of bugs in it, what is the future of the company? NONE. But, if I can identify these bugs and remove them from the code, there is a future there. Can you do that if you are not a real CFD engineer? This is now, this is not the next century, the place is one of the world leading company. As Charlie Brown used to say, GOOD GRIEF !!! We are using CFD codes with piles of bugs in it !!! The mesh independent solution? Never heard of it in industries.

These are interesting comments about job hunting in the CFD field. I personally am an Industrial Engineer being trained to do CFD with commerical codes. I will write a code for my PhD, but it will be done only so that I demonstrate my understanding of how CFD codes work and not for use in solving industrial flow fields.

Commercial codes are *believed* by managers to be the correct method by which industrial flows must be solved. I have heard supervisors comment that the CFD software companies have been doing CFD for a long time and since they spend each and every day making their codes marketable, by this reason alone, they must be doing things correctly. Unfortunately this is not always the case such as with the problem of combining a staggered grid with upwinding when a central differencing technique is not properly accounted for. However, I should hope that a situation like this would be properly accounted for in the commerical codes available.

Something important to know about industry is its time dependence on accounting. A manager follows a time table that is dependent on delivery of a product such as when the factory will spit out its product, but many managers do not like to have to deal with situations that are unpredictable. In their eyes, a commercial code is predictable. Or when it isn't, then they have someone to blame and a reason to contract with another CFD company. The products must delivered at a specified date and then the specified accounting can take place as planned so that the maximum profit is realized. This is the 'consumption machine' that we are all a part of and unfortunately, because of it, many of us must go through many layers of hoops to find a job that we are happy with.

The good news is that there are many jobs out there in which the skills we have and enjoy practicing are applicable. Unfortunately, I have yet to find a job in the field of mountain biking, rock climbing, caving or mountaineering that pays enough to feed my family and so for the time being I will have to continue being an Industrial Engineer to pay the bills. ;-)

(1). In the August 1st, 1998's posting, I had covered one area of CFD_War Principle. (2). I also mentioned issues like "license" which is required for some professional workers, like nurses. And its relations to the job. (3). I think, in order to ensure the quality of the product and the security of the job, it is necessary to establish the "exam and license" system. (4). The exam should be provided by the experienced CFD professional and a license issued by the CFD-online. Online testing can be conducted and CFD code development project can be reviewed by the experienced CFD experts. (5). I think this must be done as the first priority (6). As a requirement, in the area of using the commercial CFD code, the vendors of CFD codes must provide online training and testing to make sure that the user of the code has passed the test. (7). It is not healthy to have a online forum, job listing, discussions without any means of general quality assurance or testing for the qualification of persons and codes involved. (8). CFD codes must go through a series of testing to be qualified in certain areas and the CFD engineers must also pass certain tests and receive a license at different levels to ensure the long lasting quality of CFD work. (9). without these measures, CFD is nothing but a flea market. ( no return, no warranty)

identity is very important and its apparent that this is present in this field of research-work-thinking-beliefs. its also important that opinions have a place to reside. but its usually the users and doers who decide what to believe and in a forum such as this most would review and rethink before they embark on an endeavor such as solving a flow field numerically. because their work is for themselves, and not for someone or something else and this is a license in and of itself. however, the question is not addressed....

license cfd engineers and their codes? oh yes. license contributors to cfd-online? oh no.

Hi,

I'm a young engineer (another one looking for a CFD work). I would like to say that I completely agree with what Mr. Chien says about the "exam and license" system, particularly looking at the on-line diffusion method, which can give to nearly everibody the opportunity to submit to the exam and have some paper (or electronic reference) saing "Yes, this guy can do something in CFD", signed by some experts. Finding the desired work is a typical problem of young people, because we are not able to demonstrate (and in many cases even be sure ourselves) to be skilled enough and have the knowledge required for a particular work. That's because we don't know, in many cases, the working side of CFD, but only the academic one. So, having some experts saing that this academic knowledge is sufficient (or not!!!!) would be great for us, and, I think, for industries too. Furthermore, the exam should be something to reach, a stimulus to study and to keep update with new tendencies in CFD even if I am not still working in this field. Another very important issue is that regarding the use of commercial codes. Many job postings require the experience in using such commercial codes. As already said in this thread, one could be enormously skilled in CFD programming withouth having ever seen a commercial code. I think this one would need just a few intensive days to learn most of them, but how to demonstrate it? I think the best thing should be that commercial CFD producers distribute a demo version of their codes (widely available) so that one can learn to use them, and then access to an on-line exam which can certificate the ability in their use.

Thanks for the idea, Marco

(1). I think you are right. (2). In some well-known animation studios, they don't want you to think that the skill in commercial animation packages is important when applying for a job. They are interested in your basic training. And they also suggest places to receive training in animation if you don't already have one.( in CFD, have we defined the basic skill? ) This is because animation studios think that the basic training in animation is more important than the skill to run a computer program. In most cases, they have their own in-house animation packages and will train you when you are hired. So, the animation industry is far ahead of CFD industry in this area. (both animation studio and CFD industry use heavily the computer resources, CAD, 3-D geometry, mesh generation etc.) (3). In terms of my time on the CFD-online, I must have contributed over several thousand dollars. And I don't think Internet is the place to get free information. It is very expensive when you don't know how to erase the information you have just read.( I am talking about the accident at a school) (4). As for the real world CFD experience, most of the time, we are producing tons of output which turns out to be wrong later, because there is something wrong in the code, or in the boundary conditions or meshes. Everyday, you are facing with problems about errors made in the code by persons who are no longer with the company. It would be very frustrated to know that you have to live with the code with errors in it (bugs). For me, in the last several years, I have been using CFD codes which overwrite the user's input, produce obviously wrong results, diverge for some mesh inputs, etc... It is very interesting that people are still using the codes and trying to come up with stories and logics of how to use the wrong results in the right way. The fact is if you don't correct the errors ( for which you must have the good basic training, the know-how and the access to the source code), you are likely to receive the un-invited GPS guided surprise. (5). By the way, the school is not suppose to train a student with black box. So, you should not have the commercial codes experience when graduated from the school. (6). Under the spirit of ISO9000, an engineer must carry out engineering analysis when working on research, design and development projects. You can not ask the engineer to use a specific piece of program to do analysis, it is up to him to select the tools for his analysis work. In other words, asking engineer to use a specific piece of program to do design, development and analysis projects is not in the spirit of ISO9000 quality program. ( I remember back in early 80's, we used three CFD codes to do one analysis, and only one was able to obtain results.) (7). Most large companies have many different commercial CFD codes in their computer systems, one group would use program-A because the engineer in that group think that the answer from program-A is more correct than that from program-B. So, normally in the real world situation, you have : aero group uses programs De, Da, etc..; combustion group uses programs St, Fl, etc...; gas turbine group uses programs Ro, Ta, etc... This is a real world example. So, you are unlikely to have the time to learn all the programs in your life time. The problem is that an engineer usually got transfered from group to group. In this real world case, the only thing you can rely on is the "sound basic training and understanding" ,not the skill in using a commercial code!!! By the way, the commercial codes tend to change the contents and appearance every year for some unknown reasons. And the same program complied on workstation-A does not work on workstation-B. That is the real world.

I agree with the comments made by John C. Chien. Yes, There is no free lunch in the world. It's time for CFD people to consider the future the same as aerospace engineers in the past years. Maybe, we should develop new skills to meet the developing market though you're a Ph.D. in the area of CFD. In many times, the universities are lagging behind the society.

What you said about Universities not providing knowledge of commercial CFD codes for use in Industry (you said at least not yours) to my experience is not true. Well it is changing at least. I study at Brunel University, UK, and the lectureres, tutors are constantly looking at the commercial CFD codes available and building modules/courses that make use of them, for example to date I have roughly 4 months experience in FIDAP (4 months doenst sound long but thats when you learn most of the basics) and had to learn CFX 4.2 in order to conduct my Final Year Project and have 7/8 months experience where I learned a lot more than the basics. I am currently looking for work in the CFD area too so this line of conversation is interesting to me. I didnt mean for this reply to be a go against you, I just thought it might be enlightening for you to know that things are changing, anyway, Best of Luck Simon Assender.

(1). Yes, thing are always changing. (2). But, there is one thing which will not change, that is: some day, you will become a very good engineer, and you will be asked to do some design and analysis. How do you know that the CFD code you use will give the correct answer? And, if the design and analysis you carried out has an error in it because of the problem with the code or because of the improper use of the code, who will be responsible for that mistake? you or the software vendor? (3). That is the point I was trying to make. The point was not whether it is the ABC-university, or the D,E,F...codes. IT IS YOU !!! When a mistake is made, no one will care whether it is ABC-university, or D,E,F ...codes. ( I don't think the code vendors will guarantee that the answer you get from their codes will always correct.) At that point, you are outside the school, and you will be alone.

I agree with what you are saying, after all CFD is an approximation and that perfect answers can never be achieved. But surely with more experience and training, as Chuck was pointing out he thinks is lacking at a University level, the chances of gross error when using the code as a results of user error, are lessened? As I hope can be agreed the quality of results that are obtainable from commercial CFD codes are getting better, adding to the reasons why the majority of industry with the need for the use of CFD are turning to commercial codes as is discussed in this line of messages. Also the only real way to verify CFD predictions is to conduct accurate testing to the degree where CFD is meant to reduce....

(1). As long as you are satisfied with the tools you use, you shouldn't worry about other people's opinion at all. (2). And we all hope that this forum will be available forever to help those who need answers in the future. (3). I think, we are all aware of the situation that there are many people seeking answers in learning how to use a code. And there are also many people looking for help in knowing why the answer from a code looks strange to them. (4). The need for the CFD forum is simply because we don't know whether a code can produce a useful result or not. (5). Using CFD code is just like fighting a war. You are not going to win a war, unless you are well-prepared and well-informed. (6). As for the quality of a code, it is hard to know whether a code is improving in time or not. The code does not improve itself in time automatically. The same is true, we don't know whether the world is more peaceful or not. It is more peaceful only for those who are still alive.

(1). This is really a very good piece of information. In a way, it is consistent with my observation and experience. (2). In 90's, there is another line of jobs called " temporary subcontractor". If you are good at computer systems, programming, CAD, engineering analysis using a commercial code, this is one good approach. It is also good to accumulate experience by working with various companies. (3). It is important to know whether the company is really hireing or not. Sometimes, a massive newspaper hireing is followed by the announcement of large scale layoff. SO, the real information is very important. A friend inside can verify it quickly. Otherwise, sometimes, you will be spending several months chasing after the rainbow. (4). sometimes, it is just a routine announcement. The need may not be there. Sometimes, it is announced in an anticipation of a contract award. (5). I would say, it is important to look into other related areas of engineering. After all, not many people at human resources department understand "CFD". It is a good idea to use computer simulation, optimization, fluid dyanmics and heat transfer, aerodynamics instead of CFD, unless the job description states that explicitly. In most cases, companies are looking for people to solve problems instead of doing CFD, even if it is necessary to carry out CFD to obtain the results. (6). So, in looking for jobs, CFD is not a good key word to use.