Joel Iventosch, Computer Science graduate student at the University of Texas at Austin, spent his summer interning in the “MLP” team, WayBlazer’s R&D department. He looks back at his summer and tells about his internship experiences below.

 

What problems did you work on during your summer internship?

One of the best parts about interning at WayBlazer this summer was that I was able to work on very interesting and challenging problems. At their core, most of the challenges that WayBlazer addresses involve extracting structured data (and meaning) from unstructured data — largely from text, but also from a variety of other mediums, such as images. Specifically, a few of the challenges that I tackled during my internship at WayBlazer were:

  • Designing and implementing various models to extract structured data from free-form reviews (text) about hotels, VRBOs, cruises, and other travel related destinations or adventures
  • Contributing to the design and implementation of a Natural Language Search system, aimed at extracting structured data from unstructured natural language search queries
  • Contributing to the design of a hierarchical concept map of travel-related terms, activities, destinations, etc.

What have you learned?

I learned a ton throughout this experience, both from a technical and a leadership or “startup-ship” perspective. Specifically, I gained a much better understanding of how to apply many of the machine learning concepts that I’ve learned in academia to real-world problems in industry. A huge part of that was learning how to cope with limited labeled training data. One key takeaway was that the problems addressed in the academic setting and in industry often differ in keys ways, despite being very related.

In academia we often (although not always) are provided with a well-known and substantially large dataset and given the challenge of improving some known task or algorithm by creating innovative models or altering the algorithmic approach to the problem. In industry, however, the challenge is often the inverse: we are asked to solve specific problems related to unique data sets, using any and all algorithmic techniques available, whether they are new, old and well established, or anywhere in between — and oftentimes there is little or no labeled training data available for the specific challenge at hand, unless you or your team cultivate it on your own. Learning how to handle these types of challenges through a combination of supervised, semi-supervised, and unsupervised learning algorithms was immensely valuable.

Additionally, learning which machine learning algorithms perform best in particular settings (e.g. which algorithms perform better with limited vs. abundant annotated training data) was extremely valuable. I could go on for quite a while about everything I learned throughout my internship with WayBlazer, but this covers a few of the highlights!

How was your overall experience?

Overall, my experience interning for WayBlazer was awesome! I hesitate to even call it an “internship” experience, because it was so different from — and so far superior to — the “typical internship” experience that you hear and read about (I’ve been fortunate enough to have very good internship experiences throughout my educational career, thankfully). Most companies’ idea of an internship is to give their interns some relatively meaningless project and have them sit in the corner “staying busy” for the summer. This was the exact opposite.

The people at WayBlazer did an amazing job of making me feel like a full-on member of the team, both from the standpoint of including me in discussions about strategy, product road-map, execution, etc., and also — and as importantly — in terms of holding me accountable and responsible for getting things done. I learned a ton throughout this experience both from a technical and business perspective, and I hope that I was able to give back at least a fraction of what I gained from the experience through my work on various projects throughout the summer.

To be concrete, some of the highlights from my summer with WayBlazer were:

  • Exciting cutting edge technology projects
  • Full-on integration with the engineering team
  • The willingness of others on the team (including leadership) to listen to my ideas and consider my perspective
  • A great level of guidance and mentorship without being micro-managed at all
  • Great people and great company culture
  • Great leadership team
  • Great work environment
  • And last but certainly not least, being surrounded by very smart, talented people!

To reiterate, overall this was an outstanding experience!

What are your plans for the future?

This academic year (2016-17) I will be finishing my Masters degree in CS at the University of Texas at Austin, where I will be completing my thesis research under the supervision of Professor Scott Niekum. Upon completing my degree, I plan on entering the private sector and working for a technology company as a Machine Learning & NLP Research Engineer.

Would you recommend an internship at WayBlazer?

Absolutely! I had a phenomenal experience interning for WayBlazer and would enthusiastically recommend an internship here to anyone who might be considering it!

 

Interested in doing an internship at WayBlazer? Send us an email at research@wayblazer.com!