Getting frustrated with the living conditions in the NYC area, I'm beginning to think maybe West Coast is a better place.Is it possible for a quant developer with physics/math PhD background to make a transition to a tech company? Is math + coding skill valued in tech as much as in quant finance? Or do they prefer pure CS background? How would one go about making this happen? Would it be better to try and locate a HH first? Thanks.

It depends on the kind of tech company and the kind of product. In general, less maths and more product maintenance, especially for legacy systems.It can be a culture shock for quant developers. I would say more hardware knowledge is also needed. Industry is engineering-driven in general. More than just CS skills are needed.

Overall if you will enjoy the change as the money will be decent and even if the first move does not work out the West Coast will provide you with options. Some of the general pointers-1. Find a company which is doing "Data Science" but be vary, as it is the new sexy term, people who build charting solutions or a bit of Geospatial thrown in also call themselves "Data Science"2. Text Mining and NLP is picking up. You need to be sure that they actually do something for NLP instead of using standard packages3. "Big Data" is big these days. And moderate amount of work is related to "Data Science" and "Big Data". So look for those roles. You might want to get "Hadoop Certified". Certification matters a lot these days as there are very few people who are Hadoop capable. But all this is relevant only if you are Java guy (or C# in the slim hope that the company you apply for is looking to go with the HortonWorks version of Hadoop which is .NET enabled). 4. Lots of product companies have dedicated people for "Configuration Management" and "Release Management" work. Make sure you do not get into those roles.5. The "West Coast" folks have a lot of attitude towards "East Coast" folks. So there is a certain elitism they indulge in (of the FO BO variety). So expect some issues with fitting in initially. Eventually it will fade.Good luck!

capafan2,Is MapReduce popular?

Did interview with presumably the best name at West Cost out of curiosity. Was really surprised with mediocrity of people in comparison with tier-1 ib.

QuoteOriginally posted by: Cuchulainncapafan2,Is MapReduce popular?It is in a way and everyone is becoming Hadoop enabled. Lots of interesting stuff also happening around it. For example there is a framework called Apache Mahout which uses MapReduce to scale Machine Learning on massive datasets. But one of the limitations of MapReduce for iterative algorithms is that at each step it saves intermediate results to disk and replicates them 3 times(configurable but you need 2 atleast). Other frameworks like Apache HAMA and Apache Giraph are using the Google's BSP (Bulk Synchronous Processing) ideas. They bring messaging, barrier synchronization between steps to avoid saving to disks between each iteration. As a whole the field is promising and there isn't much outside of Hadoop MapReduce to work with. IBM's Pure Systems, EMC's HAWQ, Cloudera IMPALA and MapR's Apache Drill (based on Google's Dremel) are all tightly integrated with Hadoop.MapR is the only Hadoop implementation with strong C++ layer. I hope they provide that port shortly to bring truly high-performance computing to Hadoop. Especially CUDA integration is a pain currently due to only native Java bindings.

QuoteOriginally posted by: tu160Did interview with presumably the best name at West Cost out of curiosity. Was really surprised with mediocrity of people in comparison with tier-1 ib.IB's attract a very good technical crowd even in IT. Depending on the group of people you interview with, mediocrity can show up. Also depends on what you are looking for in a group. I know people who work in IT and are very effective at their jobs but have no idea if you asked them the difference between "Throughput and Latency". I would still have them on my team as they get things done. And I knew a guy who sat one cube away from me who could have a philosophical discussion about "Garbage Collection" but never got anything done as he loved to explore everything in depth before he got started.

Thanks for the replies. A lot of terms that capafan mentioned are completely new to me. Does that mean there is no hope for me? I mostly implement core analytics (PDE, MC, vol surface) into a C++ library, and think I am pretty good with C++ at least for the part that is needed to do my job. But, not much knowledge in hardware or network. Will I have to learn some stuff in order to make this transition? I did about 6 months of self study to jump from science to quant finance. That was possible because I had a lot of free time as a student. Not sure I can do that again now, with a full time job.

intel has also recently jumped onto the bandwagon with their own distribution of hadoop:http://hadoop.intel.com/however!OP, be wary of big data roles. although the technology is pretty cool, not every job in that space will necessarily be challenging or even interesting. a buddy of mine recently joined the data warehouse team at a company because they use hadoop extensively and he was looking to gain some exposure to it. however, 90% of his time is spent writing adhoc query reports using Pig (a dataflow language loosely akin to SQL but that relies on mapreduce to get the job done), which as you can imagine is pretty dull. so before you jump into any role being offered in this area, make sure you know what you're getting into. i am sure there are plenty of really cool roles in this space as well but just beware.you might not need to have an extensive background in hardware/networking (altho it certainly helps) because many mid-size and larger companies have dedicated syseng and techops teams to deal with the actual infrastructure+networking aspects. of course this varies from place to place and again you have to ask before joining a team what your role will entail.but yeah, there's lots of really neat stuff being done, as capafan mentioned. also look into solr and cascading if you haven't yet. google pregel is also interesting (apache giraph implements it).see: http://googleresearch.blogspot.com/2009 ... e.htmlalso, re. the certification bit - my buddy was able to get the job without being certified because he did well on his interview. so it'll help but may not be as much of a dealbreaker for firms looking to hire, if the rest of your skillset is impressive.

Any thoughts on this (and any other similar companies) with respect to OP's question? http://www.maxeler.com/about-us/careers/

QuoteOriginally posted by: ymousThanks for the replies. A lot of terms that capafan mentioned are completely new to me. Does that mean there is no hope for me? I mostly implement core analytics (PDE, MC, vol surface) into a C++ library, and think I am pretty good with C++ at least for the part that is needed to do my job. But, not much knowledge in hardware or network. Will I have to learn some stuff in order to make this transition? I did about 6 months of self study to jump from science to quant finance. That was possible because I had a lot of free time as a student. Not sure I can do that again now, with a full time job.There are plenty of roles in the West Coast who specialize in Data Science types of application where your skills will be valued. Evaluate startups carefully but worst case you will change in a year which is not such as bad deal. Java is a very valued skill in the West Coast as it becoming a very popular language for development of frameworks. LinkedIn, Google, Facebook all use Java and Python. If C++ is your forte you might want to look for firms which specialize in developing functionality which operates close to the metal. NVIDIA is one such company. Again good luck! Expect a brief rough ride but hopefully it will be worth it. The West Coast has lots of opportunity. You will find something that fits well with out skills.

Any practical tips? Should I work with a HH? Just apply directly? Thanks.

You may be able to get a job in a Financial Services firm in Bay area doing the same work what you are doing right now. Keep in mind that your Math / Physics PhD has maximum value in Financial Services and even though you have several years coding experience in C++, your knowledge level in Computer Science is actually extremely low compared to the people who tend to work at Google / Facebook / Microsoft / etc. who tend to have MS / PhD in Computer Science from top schools. The cutlure among these people and the language they speak is significantly different from what you have encountered in IB.

QuoteOriginally posted by: traderjoe1976You may be able to get a job in a Financial Services firm in Bay area doing the same work what you are doing right now. Keep in mind that your Math / Physics PhD has maximum value in Financial Services and even though you have several years coding experience in C++, your knowledge level in Computer Science is actually extremely low compared to the people who tend to work at Google / Facebook / Microsoft / etc. who tend to have MS / PhD in Computer Science from top schools. The cutlure among these people and the language they speak is significantly different from what you have encountered in IB.Google et al make market products. IB is different. People in IB have to be all-round so that they don't always have to worry about the niceties of CS, so long as it works. The trader wants it now. Microsoft can ship when they like (included their (un)documented bugs). I suppose the closest analogy to MS in IB is maintaiing a C++ pricing library.

QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: traderjoe1976You may be able to get a job in a Financial Services firm in Bay area doing the same work what you are doing right now. Keep in mind that your Math / Physics PhD has maximum value in Financial Services and even though you have several years coding experience in C++, your knowledge level in Computer Science is actually extremely low compared to the people who tend to work at Google / Facebook / Microsoft / etc. who tend to have MS / PhD in Computer Science from top schools. The cutlure among these people and the language they speak is significantly different from what you have encountered in IB.Google et al make market products. IB is different. People in IB have to be all-round so that they don't always have to worry about the niceties of CS, so long as it works. The trader wants it now. Microsoft can ship when they like (included their (un)documented bugs). I suppose the closest analogy to MS in IB is maintaiing a C++ pricing library.Microsoft is a walk in the park compared to banking in terms of stress.+1 to TJ's points of view.I would also second what tu160 said about people being not as smart. It is another way of saying what Cuch said - "people in IB have to be all-round", but on West Coast you will run into people who are far more one-dimensional, this is speaking of skills and not personality...Being a successful programmer at any west coast firm will require _significantly_ more CS than you have as a Math or Physics grad, unless you have read a lot on your own.