Today’s Example (Plus Rant): Sr. Modeler Advanced Healthcare Analytics

In my earlier post today, I mentioned, as I often do, how “recruit Americans first” policies, currently in place in the green card process and for H-1B-dependent employers, are easily circumvented, by tailoring the job ad to the foreign national whom the employer wants to sponsor. From now on, I will often post such ads, to make concrete for you readers how the system works. I’ve done so many times before, but I’ve found that many people need reinforcement. 🙂 Note that there are far more such ads than I have time to post here.

Today’s example is an ad for an opening titled Sr. Modeler Advanced Healthcare Analytics. In case you are a bit behind the times, analytics is one of the currently-favored synonyms for statistics. Other currently popular terms are predictive analytics, data science, data mining, Big Data, machine learning and so on. The first sentence of the ad is likely a good description of the job:

Designing and developing intelligent solutions working with health insurance claims data, preventing fraud, predictive modeling, statistical modeling, data mining, fraud analytics, deriving business insights, solving complex business problems, improving decisions, analyzing medical policies, risk/revenue analysis, cutting out waste and abuse…

Now, look at the laundry list of required skills:

MLR, SQL, SAS, SPSS, R Excel, univariate and multivariate regression analysis, GRM, factor analysis, sampling methods, survey sampling, clustering, classification, CART, Decision Trees, CHAID, Linear Optimization.

This is my field, folks. If you recall, in my earlier post today, I quoted immigration lawyer Joel Stewart, and noted that he has “literally written the book on the green card process” (The PERM Book). Well, I literally wrote the book on the R language (The Art of R Programming). I’ve done research on regression and classification throughout my 3+ decade career, and am currently working on a book on that too. Etc., etc. I even accidentally referred to Durbin/Grassley as “Durbin/Watson” (a statistical time series method) in a recent post to my e-mail list. 🙂

I point out my background because I “know where the bodies are buried.” Unlike the other researchers in H-1B and related issues, I know in detail what is involved in typical jobs for which the “Intels” (i.e. mainstream employers, not rent-a-programmer firms like Infosys) hire foreign students. Forgive the rant, but I have to say it’s frustrating to see analysis and policy being pontificated by people who have no idea what CART is (it’s that thing you push in the supermarket, right?) or what Linux internels are.

(Part 2 of rant.) It’s also highly frustrating to me to continue to see my foreign students get jobs from the same employers who are rejecting older Americans I know who have the same qualifications, yes, including up-to-date skills. Again, this is what the other researchers and policymakers on H-1B don’t see. I don’t begrudge the foreign students’ success in their job search, and sometimes write good letters of reference for them, but this just ain’t right.

With that point being made, let me tell you: It’s ridiculous to demand that the applicant know SPSS and SAS and R; one of the latter two is sufficient. There is similar redundancy in the rest of the laundry list.

Given how specific the skills list is, it might appear odd that the educational list is so broad: “MS in Statistics, Computer Science, or other related Quant discipline.” This is likely because the foreign national they wish to hire actually is in some “other related Quant discipline”; could be business, geography (GIS) or something like that, with a lot of coursework in “analytics.” The listed salary is suspect, likely bait-and-switch, but certainly below what a good candidate would get in an uber-expensive area like NYC.

This is a textbook study of how employers circumvent recruiting requirements, and would continue to do so if the Durbin/Grassley bill were to pass. They tailor the ad so that only the desired foreign worker qualifies. It’s the oldest trick in the book, and standard practice among the big mainstream employers in Silicon Valley and throughout the nation. I mentioned earlier the “TubeGate” videos and Stewart’s “arsenal of weapons to reject all Americans” remark, but those are only the tip of the iceberg.

The Dept. of Labor people are no dummies, and they understand that this is common. But the statutes and regulations are vague to put it mildly, lots of gray area, and in any case the DOL has no expertise in CART or Linux internals or whatever. They really have to take the employers’ word on the “need” for the stated qualifications. And need I point out that Durbin/Grassley does nothing about this? Once again, folks — business as usual for the Intels. (And for the SCEs and Disneys, as I’ve explained.)

It is of course no accident that this ad has phrasing like “Sponsorship: Yes” and “2+ years of U.S. experience.” D/G does try to address this, but really, there is nothing D/G can do about such phrasing; the ad does NOT exclude American applicants.

And last but not least, note that this “Senior” position requires only 2 years of experience. Senior! I’ve shown so many examples over the years of this, which goes directly to the connection of H-1B and age discrimination in the tech fields. Employers use H-1B as  a vehicle to hire young H-1Bs in lieu of older (35+) U.S. citizens and permanent residents.

What else can I say? Plain as day.


14 thoughts on “Today’s Example (Plus Rant): Sr. Modeler Advanced Healthcare Analytics

  1. You are unlikely to learn all those products in school, and certainly will not have significant experience with all of them in school or within two years after graduation, nor is the day long enough to exercise them all in the same day or week, or at any expert level in a month or year. To be expert at any one, probably requires a year of specialization, the polish of which wears off quickly when you switch to the next! So any laundry list like that implies a culture of mediocrity, if not outright fraud and incompetence.

    THAT SAID, such ads are not always to be taken seriously, not even as bogus qualifications that are meant to streamline a particular candidate. If you have two or three, you will be on a par with the top 10% of the candidates. And probably only one or two will actually be used in your first year on the job.

    My rant on the topic is that the whole description is absurdly hyperbolic. The typical “analytics” job is extremely mundane. If you want someone with advanced skills to truly explore masses of data for obscure insights, the laundry list of technical skills is the very least of your problems. Code monkeys seldom have much of a feel for the application domain, and a researcher with all those skills – well, you just ain’t gonna find any, people, short of the best PhDs. And presumably those insights will be worth a fortune, and you will therefore be looking to pay $500k or $2m/year or a 10% bonus on the results a ten billion dollar company gets from these insights, right? As if. But the hot trend now is to pretend you’re getting somewhere with all of these complex statistical techniques. The number of people who truly achieve it is probably a statistically close match to zero.


    • They are (usually) not hiring “code monkeys.” Some are CS majors, yes, but many are from Stat, Business and so on.

      Many people actually know R quite well by the time they get a Master’s degree, from a combination of research, courswork, work experience and so on.


      • How transferable are SPSS, SAS, S+, and R? Isn’t knowledge of the math and statistics (even, sometimes, specializations within statistics) much more important than the app you use to do the crunching.

        I knew people who, like the guy with only a hammer in his tool-kit, always seemed to apply linear programming; others who are always doing regressions and factor analysis… and then a new paper would appear in their field’s journal, and everyone in those departments would come in wanting help with another approach.

        But the different tools seemed extremely easy to adapt to.


    • What absolute nonsense. Most of these topics are exactly what you learn in a Master’s in biostat program. You learn, in most MBiost programs, SAS and R. You learn regression (MLR, regression). Now, what you probably do not learn is optimization, because those topics are from a M Operations Research. Clustering, classification, etc are from M OR. CHAID is taught almost nowhere, as it is an optimization classification. What is wierd, as Norm notes, is the broad nature of this ad. What the ad attempts to do is find the “purple squirrel”, the person who does not exist. They won’t find that person, but they can reject any person as no one can have all these skills.


        • Well, it is infamous that many H-1Bs will submit resumes that *claim* all those things, but of course it isn’t remotely true. And that is an approach non-H1Bs will also use, sometimes: claim knowledge of all things, and then paddle hard as stuff comes up you don’t know.

          But Norm, this is not necessarily the case where the foreign worker exists either. What happens is that after no verifiably purple squirrels show up, they quietly shrug and simply hire the cheapest guy on the list instead, and guess who that is.


          • In this kind of job, the foreign national that they want to hire will likely be a foreign student. It’s harder to fake things there. The student will likely claim some coursework and a research project, i.e. verifiable stuff. The student will have only minimal knowledge of some of the required skills, but this will be out in the open. As I keep pointing out, these are not “Infosys” hires.


  2. Another scam they play is this one:

    From 1988 till 2003 (maybe it has changed), we had two career paths in software.

    1. There was the microsoft camp, and
    2. There was the non-microsoft camp.

    Nothing else existed and you rarely found somebody that was proficient in both camps.

    You were either microsoft, or you were not microsoft.

    For the most part, government was non-microsoft and corporate land was microsoft land


    • And before that it was IBM (and Amdahl) amongst the B-school set, and Control Data and Digital Equipment Corp… for the scientific and engineering set. When the PeeeCeee was announced the B-school people assumed it was the only thing in micro-computing, while the artsy set drifted to Apple, and the S&E set added Data General, Sun, Silicon Graphics, Cray…

      But, mixed in there, and in professor Matloff’s last 2 postings, were people with not just technology and brand name differences in preferences, but ethical/moral differences. And none of the disputes line up neatly with political clusters (and “identity politics” hasn’t panned out in these issues, which has totally baffled many who thought that because someone lined up with particular positions on a and b that they would on c as well).

      But I agree on the point of being able to stay at the top of your game on d and e, but not on d+e+f+g… And yet, there are things that, once learned, are easily picked back up again after just a week or two. But meanwhile, 2, 4, 8, 16… (not linear, but geometric proliferation) new modules or frameworks have been launched with which it is impossible to keep up and thus avoid a lot of wheel re-invention to get the job done under pressure.


      • The excuse the employers give is that the tech sector is so competitive that they need workers who can “hit the ground running,” no learning period. This is wrong in lots of different ways, but it sells with Congress and the press.

        Liked by 1 person

        • 😎 Competitive. Yah, sure, that explains why they/we haven’t come up with any positive ground-breaking software products in a well over a decade. Oh, sure, there is a lot of derivation, and a lot of teeny tiny apps in niches already well covered. Then again, maybe I missed a couple since I’m not in the market to be buying new apps. I do see the trade pubs touting some retreads, slapping new names on old approaches, etc.

          But, yes, “hit the ground running” is an excellent pretext not to invest in training, employee relocation, etc… But it’s so obviously a flimsy pretext, who in the media would ever let the get away with it without some severe criticism? Oh, yeah, all of them.


  3. As an older programmer myself, I used to think that I could work on C++ legacy systems toward the end of my career. However, I’ve gotten the impression that even these jobs are being flooded with cheap young and/or foreign labor. I’ve considered defense jobs that require a security clearance but I don’t know that there are that many such jobs. On one of their sites, I noticed that even they ask if the candidate needs to be sponsored.


  4. I have started seeing a recent trend in employers, in an effort to “evade” potential H-1 sponsorship are explicityly saying that they require < 3 yrs experience so that they need not sponsor H-1 for a greencard (min 3 yrs experience required for EB-3 sponsorship). The skillset required is also huge. This is where the age part plays in – older folks being left out explicitly.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s