Founding father of BYOR - AI Using the BEST 2016
AI Using the Best will be the biggest online conference for data scientist, developers, tech teams and startups happening the 24th & 25th September 2016 presenting to you 100 incredible speakers by way of a novel, online conference platform. Meet Aerin Kim, Data Scientist turned Founder of startup BYOR (Construct your Own Resume) speaking at AI Together with the Best, online tech conference about her Phrase2Vec technology. Aerin is building an AI-based resume helper using NLP parsing. Each time a user uploads her resume on the webapp, it offers suggestions regarding how to enhance your resume regarding its wording or phrases.
Please inform us a little regarding your background before BYOR and the way would you get into data?
I had been a NLP data scientist in a startup called Boxfish. Used to plenty of Twitter text modeling there together been fascinated every single day from the level of information that may be gleaned from all the writing that men and women were generating. As it was a startup, our team was building the item over completely from scratch over many iterations. That training helped me later once i turned my idea into a product (BYOR).
What propelled one to push NLP parsing technology for Resumés?
My co-founder i have already been volunteering as resume reviewers and mentors for Columbia University since 2014. Yearly, we found there is a pattern for weak resumes so we found ourselves giving students the identical advice every year. We were treated to an opportunity for some automation with this resume reviewing process.
Also at school career centers, it’s challenging a one-on-one session with career advisors since the student-to-advisor ratio is hundreds to one. We made a decision to build a tool that might be employed by students to review their resume ahead of meeting their career advisors, or alternatively.
The BYOR project started because class project for the CS 224d (Dr. Richard Socher) at Stanford. Rohit and that i took that class online.
How do you train the term embedding neural networks to locate similarities and relations between phrases?
The primary strategy for finding similarities and relations between two different phrases is converting these phones phrase vectors and then seeking the distance between these vectors. There are numerous ways to calculate phrase vectors. The best way that one can try is usually to first train the phrase vectors then weight average those word vectors utilized in the phrases.
So what can BYOR do when compared with other CV checkers?
Currently, there isn't any company that means result phrases on the specific sentence. Even AI companies with good level of funding don’t open their platforms like us. Inviting visitors to upload virtually any resume and provide them suggestions can be a challenging problem on many levels and taking it on requires a little bravery.
What traditional CV checkers do is simple keyword extraction or keyword counting to evaluate whether certain language is used you aren't. They don’t see the user’s resume line by line semantically.
What’s been one of the most exciting a part of your startup adventure?
One of the most exciting part happens when we enhance the “phrase suggestion algorithm” day-to-day and flourish in generating phrases that produce sense.
Also, ahead of the startup, I did previously help a huge bank. If you are a employee of a giant company, your job description is quite narrowly focused. But also in a startup, I'm able to try out every aspect from the product. Many experts have extreme fun for me so far.
Also, it’s amazing to see lots of people causing BYOR voluntarily.
If it’s not a secret, that's your favourite technological setup?
It’s a well known fact. We use python django for web. All NLP/deep learning code is written in python.
To train word vectors, we use code developed in C.
What advice can you share with budding AI developers?
If you're AI developer, Applied Math basics are necessary to suit your needs. Invest a number of your time to debate Linear Algebra, Optimization, Probability that you just learned during college.
Are you currently pumped up about speaking at AI Using the Best?
Yes! I prefer that it’s priced under 100 bucks to ensure average person can attend. And it’s on the web!!! People/students shouldn’t have to have sponsors to attend such tech conferences. With The Best line-up can be as good being a $3000 conference.