So far I've heard back from David Hardtke and Emily Thompson with follow up information. I will update this with information from others when I get it. All of the panelists were willing to share their email addresses, but I have not typed them in here without their explicit permission to post it publicly. Emily Thompson: emily@insightdatascience.com, https://www.linkedin.com/in/enthompson David Hardtke: dhardtke@pandora.com The other panelists typed them into the panel during the session, but if you were not able to write them down, contact Christine.Nattrass@utk.edu. David Hardtke noted that this is application season for entry level positions and that he has an opening on his team. Names of programs and resources referenced during the panel: Programming languages: Python Emily Thompson: For other resources, the book I mentioned by Joel Grus called "Data Science from Scratch" is an O'Reily book that does a nice breadth-first approach to all things data science. There's a lot of good lingo in there for people who are lost in all the things they might want to brush up on before interviews. We also have a blog here that helps with the kinds of courses, tools and tutorials that can help get academics started: https://blog.insightdatascience.com/preparing-for-the-transition-to-data-science-e9194c90b42c#.64hyxz7vq I also wrote a blog post about some FAQ that academics often have about transitioning to data science: https://blog.insightdatascience.com/academia-to-industry-data-science-myths-and-truths-387e1e07dec0#.i3yri9irz Finally, we're accepting applications now for our January session of Insight Data Science, with a deadline of October 24, 2016. Anyone with a PhD in physics is welcome to apply! Find out more on our website: insightdatascience.com David Hardtke: http://cs229.stanford.edu/ This is the standard baseline set of Machine Learning Knowledge with “correct” terminology https://www.glassdoor.com/Interview/data-scientist-interview-questions-SRCH_KO0,14.htm Most companies use the same interview questions and many are on glassdoor Sanjay Arora An overview of the finance industry: Traditionally, most physicists transitioned from academia to finance. One can split finance into two broad categories: banks ("sell-side") and hedge-funds ("buy-side"). Banks (Goldman Sachs, JP Morgan Chase, Morgan Stanley, RBS, etc.) theoretically exist to provide "liquidity" to markets i.e. to assign capital i.e. money from areas where demand for it is low to areas where demand for money is high. After 2008, many of the banks transitioned to being "market-makers". This means that you can call up any bank and ask to buy/sell a "security" (stocks, bonds, derivatives like options, futures) and they are legally obliged to follow through. They are free to quote any price they like though and if both parties agree, the transaction goes through. This means that the main problem at banks is managing "risk". They don't want to be stuck with assets/securities that will sharply lose value and they have to carefully manage what they hold. Increasingly strict regulations are also a major challenge for banks. The US government requires that each bank holds enough capital "on the side" in case it goes broke. In these cases, the banks can bail themselves out unlike 2008, where the government had to step in. The banks want to set aside as little capital as possible because free capital is a "wasted" asset. Capital is useful when it is earning interest and invested in securities. From the government's perspective, each bank should hold a large enough amount of capital that it can bail itself out. This natural tension plays out in regulation after regulation and banks do a lot of modeling ("stress tests") to show that even if the markets collapse, they have enough capital set aside. Even today, when most physicists pick up books about finance, they focus on pricing derivatives (contracts involving stocks, bonds etc.) using tools like stochastic calculus. The reality is that banks no longer create exotic derivatives and price them. Those activities were mostly regulated away. Instead, the big problem is risk management. This involves a lot of data analysis as well as understanding the markets well. A last point about banks. They generally have the following abstract structures/divisions: Sales and Trading: Where trading actually goes on. This houses the market-maker side too Investment banking: helping companies go public, raising capital. very relationship-heavy business Investment management: managing wealth for rich clients including individuals. Ideas like portfolio optimization i.e. how should iinvest your money to hit a targeted return rate with acceptable risk Technology: doesn't make money so second-class citizen. Still crucial because technology distinguishes execution capabilities of one bank vs another. Compliance: making sure bank is compliant with laws Hedge funds are at the opposite end of the spectrum. They tend to be much smaller. Some super-successful hedge funds have < 200 people. They exist for one purpose - to make as much money as possible using any (legal) means they can. Prominent examples are Renaissance Technologies (full of physicists and mathematicians), Citadel, Two Sigma (very "quanty/techy"), Hutchin Hill (more like Renaissance), Blue Mountain etc. New York has thousands of hedge funds and they are usually very secretive and keep a low profile. They take money from investors and get to work. Many have a research environment where exploring new technologies, mathematical techniques, market ideas is heavily encouraged. It's generally hard to start at most of them straight out of grad school/academia because they look for experienced people. While one might have ethical concerns about purely making money, the work environment at a hedge fund can be intellectually very stimulating. An overview of the tech industry: One can divide Silicon Valley (and extensions in New York, Seattle etc.) roughly in the following groups: Big established companies that are seen as being central: Google, Facebook, Amazon, Microsoft (to a lesser degree), Netflix Old companies that used to be central: Oracle, Cisco, IBM Startups that are big and considered successful and where many people want to work (specially before they go public): Uber Airbnb Palantir SpaceX Snapchat (dozens more) Then there are the startups that are small and struggling but super-exciting to be at and where one gets to learn an enormous amount. These are peppered everywhere in silicon valley. For a physicist, spending a couple of years at any of these would be an enormous learning experience. While work at a place like Google can get mind-numbingly boring given its size, it's probably one of the best training grounds for a new software engineer. Learning how to work with giant code bases, writing tests, using a repository and then diving in and learning how build systems work or transaction systems work - all of this is technically very interesting and makes one a much better engineer. As a data scientist, my impression is that work at Google and Facebook is mostly related to serving advertisements. If that excites you, go for it. On the other hand, the startups can give you access to many more interesting data problems. Warning: don't get taken in by the public image of these companies. Most engineers at google don't work on self-driving cars and deep learning. But, don't let it discourage you either. Go with an open mind and see what you can learn and what you find interesting. Work environment: IMHO, there are a few things one should pay attention to in any job: Learning: How much and how fast do you get to learn? What do you get to learn - technical software engineering, machine learning, more general data science/statistics, how to build a product and take it to market? What do you care about learning? "Passion": Do you care about what the company does? Do you care about what your day-to-day is? It's okay to be neutral despite all the talk about "finding your passion". It's also okay to go through phases where you go from loving your work to disliking it (if you really dislike or hate it, it's time to switch jobs). Overall, caring about something makes it much easier to go through rough patches. Money: How much do you care about it? Are you happy with a good salary? Do you want to make a lot of money and are willing to sacrifice quality of work? Only you can make that decision. The only observation I have is that mostly doing what one loves leads to making more than enough money, both in finance and tech. One warning: almost everyone says they won't let money affect them but everyone eventually falls prey to it. Be careful of this because it can lead you to a career path where compensation outpaces your skill set which puts a big target sign on your back. Office environment: This is interesting. My first job at Goldman required wearing a suit, showing up at 7:30am and staying there till 7-7:30pm. It also required being at my desk on the trading floor most of the time which meant all lunches were eaten at the desk. I also ended up working most Saturdays and Sundays. To top it off, I was not doing anything mathematical or writing lots of code (mostly just simple scripting). Eventually I got frustrated and joined another team where my work involved lots of machine learning and coding, there was no work during weekends so I could work on personal projects and the environment was far more relaxed. I used to go at 8am because I was excited to start work and I left at 6-6:30pm but again because I wanted to work. If I needed to leave at 3pm, I could just tell my team-members and leave. And the best part - no suit! Banks are very formal and time-intensive in front-line divisions i..e parts of the company that make money. Technology groups at banks are far more relaxed and the hours are very reasonable and even flexible sometimes. Hedge funds tend to be even more relaxed because they have no pressure of quarterly reporting or regulation and exist to do research and make money. Tech companies take this to another level. As long as work gets done, most companies don't care when you come to the office and when you leave. It's preferable to spend at least a few hours with your team-mates most days but if you need to focus and work from home, no one even notices. The dress code is super-casual. I work with colleagues who come in pajamas sometimes. There are free meals at the office and people without families tend to eat dinner at the office. Many people prefer keeping evenings separate from the office and go home. Anything goes as long as work gets done which is very liberating. Resources: For interviewing, I would ignore most of the below except for hacker rank, a couple of coursera courses and then start real interviewing. It's easy to get overwhelmed so focus is essential. Github: https://github.com/ There's no rocket science here but many companies specially startups look for an applicant's github account and the code they have. Don't worry if you have nothing on yours. Keep applying. I still don't have any code on my github. Kaggle: kaggle.com Kaggle is a site where companies post data science problems. It's a good way to pick up some machine learning techniques and to play with new datasets. Many employers discount kaggle experience though. Some reasons are: 1) kaggle datasets are super-clean compared to real-life datasets, 2) the aim is to get the highest possible score (according to some metric) and this results in very complicated models which are combined to form even more complex models. No one needs to do this in real-life and it's not valuable to build a model that takes some error from 0.00024 to 0.000235, 3) the datasets are generally small and fit in memory which is not the case in many real-life problems (although I still think the vast majority of problems do fit in memory). But, if you have never tried machine learning models before, it's a very fast way to get up there. Glassdoor: https://www.glassdoor.com/index.htm See what interviews are like at various companies Hackerrank: https://www.hackerrank.com/ Practice coding problems. This is extremely useful for interviews. Probably much more important than github or kaggle. Free online courses at Coursera, Udacity, Edx: This can be overwhelming given how many courses there are. I finished two in the last 6 months of grad school, mostly 4-5 hours each weekend. Algorithms: https://www.coursera.org/learn/algorithm-design-analysis Machine Learning: https://www.coursera.org/learn/machine-learning I learned a lot from both of them and they demystified a lot of concepts. Overall, most coding interviews will involve hands-on coding (hacker rank) and a knowledge of algorithms and running times (the coursera course). If you are interviewing for a data science position, the coursera course in invaluable. If you are interviewing for finance quant roles, there'll mostly be questions on algorithms and mathematical problems but not much finance. To get an overview of financial concepts, the following two are useful (don't try reading everything!): Options, Futures and Other Derivatives - not very mathematical but very good explanations https://www.amazon.com/gp/product/0136015867/ref=pd_sbs_14_t_1?ie=UTF8&psc=1&refRID=V0MMEMWW9TWVW03NAHHJ Books by Paul Wilmott https://www.amazon.com/Paul-Wilmott-Introduces-Quantitative-Finance/dp/0470319585/ref=sr_1_1?s=books&ie=UTF8&qid=1475678740&sr=1-1&keywords=wilmott+quantitative+finance Warning: Please don't try reading every chapter from any of these. Even just knowing something about options is mostly enough. A lot of the later chapters are not even used in the form they are described. Machine Learning (more long-term if you decide to continue in the field): Some are free online so please google https://www.amazon.com/Pattern-Recognition-Learning-Information-Statistics/dp/0387310738/ref=sr_1_1?s=books&ie=UTF8&qid=1475678851&sr=1-1&keywords=christopher+bishop https://www.amazon.com/Machine-Learning-Probabilistic-Perspective-Computation/dp/0262018020/ref=sr_1_1?s=books&ie=UTF8&qid=1475678866&sr=1-1&keywords=kevin+murphy+machine+learning https://www.amazon.com/Bayesian-Reasoning-Machine-Learning-Barber/dp/0521518148/ref=sr_1_1?s=books&ie=UTF8&qid=1475678893&sr=1-1&keywords=david+barber Pick one (i think the last one is free) to get a nice overview during your first job. Then there are many more focused books on bayesian learning, neural networks, information theory, kernel methods etc. If you find core computer science interesting: Computer architecture: https://www.amazon.com/Computer-Architecture-Fifth-Quantitative-Approach/dp/012383872X/ref=sr_1_1?s=books&ie=UTF8&qid=1475678976&sr=1-1&keywords=hennessy+patterson An article by my old boss and mentor about memory: http://david.jobet.free.fr/wiclear-blog/images/cpumemory.pdf