jayshahtx

This is true, but in a lot of instances ignoring variable dependence still produces meaningful results. For example, Naive Bayes on text classification usually performs really well even though words occurrences are dependent on each other. From my experience, if you are attempting to classify on a large number of observations (E.g. a large number of words), the algorithm performs well.

It's nice to see this make the headlines, I was at Microsoft M&A group in 2012 when the acquisition became public. Satya has received a lot of praise for the direction Microsoft has been going in, but I still feel that some of the groundwork was laid by Steve. PPI is a good example with the acquisition becoming public in summer of 2012.

When most people today talk about Steve, they often make it sound like he was aloof to the needs of the industry, the fact that Microsoft needed to be re-organized, etc. I don't feel like he didn't know these things, but more-so that it took a change of leadership to rally the company with different goals. [edited to add more commentary]

I worked for a very early stage startup that ended up getting press from WSJ, TechCrunch, etc but I had no visibility into the process. Can someone who has received similar press comment on how they managed it? Is it similar to OP's article or did you have connections?

I agree with you, but when we are discussing "price", we are not discussing it as a function of profile. We are discussing it as a function of raw skill. If two people (one experienced and one in-experienced) were asked for a proposal on the same specs, the experienced one would demand a much higher rate, even if the final output would be the exact same. Using the word "underpriced" from the lens of competency makes sense. Your ability to gain trust/respect from a client? Different story..

> pass on the cheap, low-value projects. Hold out for the clients that have budgets and value your time. Every second you spend with clients that don’t have money and don’t value your time is another second you have wasted.

I just finished my first 8 months of consulting work and while this is true, it is so difficult to actualize. When we had zero projects, I would take any meetings I could get. A part of me would hope that during the meeting, the client would realize where I could add value and increase their willingness to pay. In reality, however, it was the client who tried to push down our rate time after time. A hard lesson for a novice, but I lost hundreds of hours to clients that I knew, deep down, didn't have the budget.

With that being said, this is dangerous advice to internalize pre-maturely. Without underpricing yourself in the beginning, it's difficult to derive self-worth (IE: confidence) and increase your rate later on. And the last thing you want is an early freelancer to turn down work because they think they are worth more.

TLDR; Underpaid work might just be a rite of passage in freelancing. It was for me.

Loved this and really like Madeon - also worth checking out the Daft Punk equivalent to this/very similar concept:

http://www.najle.com/idaft/

This is an excellent article - many of my mentors have guided me not to pursue funding for as long as possible, until I'm cash flow positive, etc and its great to see a discussion on HN reinforcing this sentiment. In a day and age when so much can be accomplished with just a laptop (git push heroku master), I am surprised bootstrapping is not a bigger topic of conversation.

This reminds me of a study I did on CEOs who came to speak to my university. I found that investing in a company if their CEO was visiting college campuses averaged an annual return of ~20%. The logic was that CEOs of poor performing companies didn't have time to visit schools.

For a disclaimer, the portfolio was also more volatile than most benchmark indexes, so looking at the Sharpe ratio in both instances (mine and OP's) is important.

Hey Caser,

Very interested in applying - are you still accepting applications? I saw the first trip starts this month.

This is very similar, if not the same, technology as Husk Power Systems. HPS had a lot of press around 5 years ago for going to villages in India which grew corn and burning their husk for power using tech from WWII. They were so successful that corn husk became a commodity in the villages they entered. A part of their model was also to train locals to become engineers and manage the machines themselves. Pretty incredible value add at many different levels.

http://www.huskpowersystems.com

I think the most interesting datasets are within reach but require curation yourself. For example there are extremely powerful scraping libraries in just about every popular language today, not to mention APIs such as Twitter's.

If you're looking for a cool dataset to play with, I think it is more productive to ask yourself what questions you want to answer and then find/curate the data VS find a dataset and then ask "what questions can I answer?". The former approach will also keep motivations high if you're driven by curiosity.

SEEKING WORK

Location: Austin, TX

Remote: Yes

Willing to relocate: Temporarily

Technologies: Full stack, Machine learning, data science, Angular, AWS, Google App Engine, Python

Résumé/CV: www.linkedin.com/pub/jay-shah/30/aaa/926/

Email: jay[at]bayesconsulting[dot]io

We're a two man machine learning shop and help companies build new solutions with old data. We've automated ad copy for Google PLAs, helped a staffing company parse 100K+ resumes with NLP, and are helping a local oil and gas company predict oil production. We love hard problems and breath data science. Recently finished with CS degrees in a top 10 CS uni with 3.97+ GPAs. Please feel free to send an email!

  Location: Austin, TX
  Remote: Yes
  Willing to relocate: Temporarily
  Technologies: Machine learning, data science, Angular, AWS, Google App Engine, Python, (Full stack)
  Résumé/CV: www.linkedin.com/pub/jay-shah/30/aaa/926/
  Email: jay[at]bayesconsulting[dot]io

We're a two man machine learning shop and help companies build new solutions with old data. We've automated ad copy for Google PLAs, helped a staffing company parse 100K+ resumes with NLP, and are helping a local oil and gas company predict oil production. We love hard problems and breath data science. Recently finished with CS degrees in a top 10 CS uni with 3.97+ GPAs. Please feel free to send an email!

I've thought about this topic quite a bit, and I don't think the root emotion in many "mean" people is, well...meanness. I think it is a strong desire to control. This desire stems from the importance of managing execution.

To be quite honest, if Steve Jobs wasn't dominating, maybe the iPod Nano would have been much thicker. Maybe the Facebook experience would have sucked. Maybe Tesla would have filed bankruptcy. There is a long list of "Maybes" which could go here. From a societal standpoint, it's interesting. Ask someone in 1991 if Jobs in an asshole, the resounding answer is likely to be "yes". Ask someone today, and it's almost always pardoned with a "but he was a genius".

I had this very internal debate with myself throughout undergrad, particularly when you are placed in a group where you care the most for what ever project. The one who cares the most tends to have the most complete vision, and thus will strongly desire the pieces to fall a certain way.

I'd like to believe that not everyone is mean, but I can't help but wonder how possible it is to avoid "meanness" when the success of your organization hinges on your ability to execute a vision no one else can realize.

This caused the spinning wheel of death on my maxed out MBP...am I alone?

I wrote my senior thesis on a calendar assistant that achieves much of the functionality mentioned here (recommending times/places, auto updates, etc) and lost so much motivation just to the sheer amount of integration between parties/services required to solve this problem. There has been a lot of work in the academic space on optimizing calendar/email functions but its challenging to implement across so many email providers/clients. I'm glad Google is taking a step in this direction, I'm not confident many startups would be able to solve this problem effectively.

Am I the only one who doesn't think $11M is really going to move the needle for SF OR AirBnB?

Location: Austin, TX

Remote: Yes, have experience

Willing to relocate: In Spring 2015, yes

Technologies: machine learning/information retrieval methods, java, python, matlab

Resume: http://linkd.in/1uo7yLq, www.jayshah.me

Email: jayshahtx@gmail.com

Hire me to analyze/curate large data sets, extract insights, and build predictive models.

Previous work:

- Used machine learning to automate $1MM service at Umbel (Austin Startup)- recommended ads for clients to run to Facebook by analyzing 1B+ data points

- Used local Twitter sentiment to predict quality of life (model used to predict poverty rates, population density, etc)

- [Current] Using machine learning forecast concert ticket prices

I'm a triple major student at UT Austin with an almost perfect GPA. Graduating this year, lover of all data. Email me at jayshahtx [at] gmail dot com or visit me at www.jayshah.me

I found this comment to have an undertone of ignorance.

1) Implying a cluster of 54 countries are all homogeneous, have poor health standard, etc

2) "Struggles to provide any clean water to humans at all" - this is much more heresy than fact. Please rely on statistics

African countries by no means are world superpowers, but this comment is pretty ignorant. Please think twice about such comments, they can be interpreted as offensive

What a great website with an intuitive product. Cherry on top is the extremely relevant blog post which is A) somewhat interesting B) obviously connected to their product C) not regurgitated industry information.

This might not go viral, but this is what I define as good marketing/messaging efforts

Side note: I'm always a little bothered when people attribute something from "Africa". Africa has 54 countries, c'mon people

I'm always surprised by Amazon's lack of fanfare. Even for such a huge release, their product page is literally the same as that of something as random as a blanket

Link-bait is the most frustrating trend on the internet. I read somewhere when outlets such as NYTimes were interviewed on their perception of BuzzFeed. The overwhelming response was "we're not worried at all, BuzzFeed has not taken any traffic from us".

On a similar note, clicking on links that take you to a YouTube video vs just playing the YouTube video (embedded) is immensely annoying.

Interesting idea, but wasn't able to find much literature on this on the internet. Are all the proper nouns used in your post hypothetical? Apologize if I missed a clear hypothetical...

Would love to see a few basic libraries install ahead of time (NumPy and SciPy, namely) and hear about performance implications of running virtually.

Cool tool though - I've done data science work for a few companies now and the most frustrating thing is always getting set up on their stack.

Hire me to analyze/curate large data sets, extract insights, and build predictive models.

Previous work:

- Used machine learning to automate $1MM service at Umbel (Austin Startup)- recommended ads for clients to run to Facebook by analyzing 1B+ data points

- Used local Twitter sentiment to predict quality of life (model used to predict poverty rates, population density, etc)

- [Current] Using machine learning forecast concert ticket prices

I'm a triple major student at UT Austin with an almost perfect GPA. Graduating this year, lover of all data. Email me at jayshahtx [at] gmail dot com or visit me at www.jayshah.me

[Location] - remote or Austin, TX

[Technologies] - Information Retrieval, Machine Learning, general data science, most work in Python

[Contact] - jayshahtx[at]gmail[dot]com

www.Fortune.com just moved to this menu style

SEEKING WORK - remote (based in Austin)

Hire me to analyze/curate large data sets, extract insights, and build predictive models. Significant experience in Information Retrieval, Machine Learning, general data science

Previous work:

- Used machine learning to automate $1MM service at Umbel (Austin Startup)- recommended ads for clients to run o Facebook by analyzing 1B+ data points, forecast ad prices with 98% accuracy

- Used local Twitter sentiment to predict quality of life (model used to predict poverty rates, population density, etc)

- [Current] Using machine learning forecast concert ticket prices

I'm a triple major student at UT Austin with an almost perfect GPA. Graduating this year, lover of all data. Email me at jayshahtx [at] gmail dot com or visit me at www.jayshah.me

Machine learning focused CS undergrad in Austin, TX - would love to talk to a data scientist in the industry for 30 mins.

I sense there is a large disconnect between ML research at university and application of ML in industry. I'm most curious how teams work at Netflix, Palantir, etc (at that scale and company size). I perpetually feel the impostor syndrome in this space (even though I performed very well in CS course work and research). If you can give me a glimpse of data science at your company, I would be grateful!

jayshahtx[at]gmail.com

SEEKING WORK - remote (based in Austin)

Hire me to analyze/curate large data sets, extract insights, and build predictive models. Significant experience in Information Retrieval, Machine Learning, general data science

Previous work:

- Used machine learning to automate $1MM service at Umbel (Austin Startup)- recommended ads for clients to run o Facebook by analyzing 1B+ data points

- Used local Twitter sentiment to predict quality of life (model used to predict poverty rates, population density, etc)

- [Current] Using machine learning forecast concert ticket prices

I'm a triple major student at UT Austin with an almost perfect GPA. Graduating this year, lover of all data. Email me at jayshahtx [at] gmail dot com or visit me at www.jayshah.me