AI With The Best may be the biggest online conference for data scientist, developers, tech teams and startups happening the 24th & 25th September 2016 providing you with 100 incredible speakers by having a novel, online conference platform. Meet Aerin Kim, Data Scientist turned Founder of startup BYOR (Make your Own Resume) speaking at AI With all the Best, online tech conference about her Phrase2Vec technology. Aerin is building an AI-based resume helper using NLP parsing. When a user uploads her resume around the webapp, it offers suggestions on how to improve your resume regarding its wording or phrases.
Please tell us a little concerning your background ahead of BYOR and the way do you get into data?
I used to be a NLP data scientist at the startup called Boxfish. I did plenty of Twitter text modeling there and had been fascinated every single day with the level of information that may be gleaned from all of the writing that folks were generating. Given it would have been a startup, all of us was building the merchandise on your own over many iterations. That training reduced the problem later when I turned my idea into a product (BYOR).
What propelled that you push NLP parsing technology for Resum�s?
My co-founder i are already volunteering as resume reviewers and mentors for Columbia University since 2014. Every year, we found you will find there's pattern for weak resumes and that we found ourselves giving students the identical advice year after year. We got a chance for some automation within this resume reviewing process.
Also at school career centers, it�s difficult to get a one-on-one session with career advisors for the reason that student-to-advisor ratio is hundreds to at least one. We decided to build a tool that might be utilized by students to examine their resume just before meeting their career advisors, or alternatively.
The BYOR project started as the class task for the CS 224d (Dr. Richard Socher) at Stanford. Rohit and i also took that class online.
How will you train the word embedding neural networks to find similarities and relations between phrases?
The key way to find similarities and relations between two different phrases is converting them to phrase vectors and then locating the distance between these vectors. There are numerous ways to calculate phrase vectors. The most effective way that one can try is to first train the saying vectors then weight average those word vectors employed in the phrases.
Exactly what do BYOR do in comparison with other CV checkers?
Currently, there is absolutely no company that implies result phrases on the specific sentence. Even AI companies with higher volume of funding don�t open their platforms like us. Inviting people to upload just about any resume and provide them suggestions is often a challenging problem on many levels and taking it on needs a little bravery.
What traditional CV checkers do is not hard keyword extraction or keyword counting to test whether certain words are used or otherwise not. They don�t comprehend the user�s resume line by line semantically.
What�s been one of the most exciting part of your startup adventure?
Probably the most exciting part happens when we enhance the �phrase suggestion algorithm� everyday and flourish in generating phrases that make sense.
Also, prior to startup, I used to benefit a huge bank. If you are an employee of a giant company, your career description is quite narrowly focused. In a startup, I will experiment with every part of the product. It's been exciting for me so far.
Also, it�s amazing to see a lot of people adding to BYOR voluntarily.
If it�s a well known fact, which can be your favourite technological setup?
It�s not a secret. We use python django for web. All NLP/deep learning code is written in python.
To coach word vectors, we use code developed in C.
What advice can you give budding AI developers?
If you are AI developer, Applied Math basics are necessary for you personally. Invest some of your time and energy to talk about Linear Algebra, Optimization, Probability that you simply learned during college.
Do you think you're looking forward to speaking at AI Using the Best?
Yes! I love that it�s priced under 100 bucks to ensure that general public can attend. And it�s on the net!!! People/students shouldn�t need to have sponsors to attend such tech conferences. Together with the Best line-up is really as good like a $3000 conference.