Are you looking for an academic position?

How does one give a good academic job talk? I struggled unsuccessfully with this during my days in the US, and eventually returned to India with a job I got more easily. It worked out for me. But the question still needs an answer.

A senior professor once told me, "80% of them should follow 20% of your talk, 20% of them should follow 80%, and 0% of them should follow 100%."

That's an amusing but useful principle. How does one realize such an ideal? Many years after I needed it, I eventually came up with a cartoon of such a talk. Luckily I have never had to give such a talk, and I doubt if I could give one. But here is the cartoon.

(a) Begin with a big number they understand and accept, but preferably did not think about before. Like there are 2.3 billion apple trees in the US. Or every year, long distance flying consumes 163 cubic miles of aviation fuel. (I made those numbers up, you would use actual ones.)

(b) Point to an improvement or efficiency or some other thing with obvious impact. Saved money or saved lives, both are good. So, we could save 1.3 million tons of apples, or 12 cubic miles of fuel. Suggest gently that you might be able to achieve 1-3% of that. Convert to money, but don't dwell on it. People have fast eyes for money, they'll see it without trouble.

(c) Jump suddenly into a technical problem. Like, "what is the deformation mechanics of fresh fruit?" Or, "what are the issues involved in heat transfer and flow near a turbine blade?" Give them a mild surprise and show them it is not all fun and games.

(d) Show 2 or 3 distinct pillars of the problem area. Computational challenges, state of the art, groups in big universities that you are in touch with. Remarkable experimental setups: "they can hear a spider web tear", or "they can extract stress strain relations from a specimen the size of a sand grain". Theory, modeling, basic principles, a slide with a few equations with condensed notation which uses special fonts, "that's energy balance and exchange between species in the flow" and "that's the continuum irreversibility, diffusing as it moves downstream" (I am making this stuff up). Here you are showing them what a big problem it is that you are taking on, and how you understand the work of the biggest people of this field (and are maybe even friends with them).

(e) Identify a subproblem within one pillar. Show depth and breadth within that. Show that you know details of it that the audience does not know. Here you are giving a glimpse of realistic hard labor done by an extremely competent person (yourself).

(f) Suddenly, make a shift. Say, "Here's a simplified and idealized problem." Spherical, or homogeneous, or one-dimensional, or with a single purely cubic nonlinearity ... depending on what you know and can do. Point out that it connects with something Pascal wrote in a letter to somebody, or Euler wrote on the back of an envelope. Show how that great man's comment leads to a dramatic simplification of your problem, leading to a difficult-looking integral. Compute the integral, preferably as an asymptotic approximation, and give an answer, preferably including the square root of pi. Show a little helpless joy.

(g) Sigh. Say, "OK, let's get back to the more complex problem." Now present some results. Graphical is good. Maybe complex and colored, based on difficult work the public can respect, but with interpretations that the public can also understand quickly and easily, like "the shock is moving up", or "captured into resonance here for a while" or "this edge vortex is the main culprit" or "each reflection cuts the energy in half" ... Here you are bringing the lost ones back in by letting them feel engaged (and smart).

(h) Get ready to wind up. Point to how much still remains to be done. How this is a rich area they should invest in (by hiring you).

(i) Conclude. Lightly remind them of the bigness of the problem, but move quickly to the broad scope, the specific area you will work in, and how there remains so very much to be done. Here some influential sleeping people are waking up, and you need to get them back into the group.

(j) Remember, when different types of people agree that you should be hired, they don't compare reasons.

---------------------------------------------

Good luck. Since I cannot give such a talk myself, I do not have a video to share.