Loading...

Software & Machine Learning Engineer

Hey, I'm Oktay

About

Get to know me
Team Member

Oktay Bahceci

Software & Machine Learning Engineer

About

Hey! I'm Oktay.

I'm a Software and Machine Learning Engineer, coder & entrepreneur with a M.Sc in Computer Science with specialization in Machine Learning. I love to solve problems and create new stuff with whatever tools available. I've been in the creating and optimization game since I was a kid, and I like investing myself into side projects. I like to feed my creative side once in a while to create artsy things on my free time.

> I'm looking for new challenges.

Feel free to contact me, I'd love to connect!

Interests

Coding

I love coding. Learning new languages & tools to finish a thought or project is a part of the workflow. It's such an amazing and liberating feeling to be able to create whatever I can think of with my mind and fingers. It's amazing that developing and testing the end product can be done for free.

Artificial Intelligence & Machine Learning

I've always been curious and a digger. My curiousity makes me dig until I feel like I have a bird's eye view of something. Then, stats came along and changed everything. Machine Learning made finding, predicting and analysing so much easier. I've specialized in ML for three years now, mostly focusing on neural networks. Making predictions & building intelligent agents that take use of statistical estimation and optimization makes me excited.

Biotech & Bioinformatics

I took a class in bioinformatics and instantly fell in love with it. Using DNA data to understand information processes of nature & living entities is pretty damn cool and has limitless applications. In the future, I would like to help push the applications of ML in biotech so more good can be done.

Mechanics & Design

After analysing behaviours and information processes, I have started to take interest in complex mechanics. I got some experience of creating complex design models in CAD & got to physically create them in high school. I've become fascinated by Mechanics & Design and would love to learn more.

timeline

from Script Kiddie to Master

Timeline

A brief history



  • 1997

    Learning DOS and C

    The image above is of me, age 4, soon after my mom purchased a Windows 95 system, and right before it changed my world. I was lucky enough to have my curiousity nurtured with jigsaw puzzles, problem solving toys, kinder eggs and I was reading by the age of 3. With this system however, I had a completely new thing in front of me to feed my curiousity. Little did I know that it was going to give me the superpowers I had always wanted.
    Anyone wanting to hire a code wizard?

  • 1998

    Creating custom game levels

    I explored every bit of Windows and DOS there was. Games was fun, and I wanted to create my own. Could I combine the commands I learned and create my own puzzle? After exploration, I found the files responsible for the levels, analyzed them, copy pasted around, and eventually, I had created my own game levels.
    I created something, and it felt great.

  • 1998-2000

    Other computers & Internet

    I started first grade in 2000 & we had one computer in the classroom. I swept through classwork to get computer time. Internet started to become available, but was very limited. After exploring the software offline, there was a way to connect and find more information.

  • 2000

    Script Kiddie

    When I was in first grade, a teacher asked me if I wanted to be a "hacker", and boy, they shouldn't have. I Altavista that and found a ton of C code. I low key became a script kiddie. I was the coolest kid on the block with my floppy's containing system calls & while loops that could do a lot of damage. I still remember seeing the console countdown before seeing computers being toasted.

  • 2000-2005

    Web Designer & Photoshop

    By this point, I felt like a computer expert. I could reformat any computer, install and download whatever software or game I wanted, without spending mom's money. Swedish social networks such as Lunarstorm and Playahead arised, and you could do so many cool things with your Playahead profile. It was simple - you just changed the HTML, CSS & used Javascript to add a song to your profile, and used Photoshop to create cool images. If I needed to know how to do something, I just had to Google it.

  • 2005-2009

    Web Developer & Hacker

    In 2005, I pretty much hacked during the nights when I was bored. I took every design, web development and coding job I could get, primarily using PHP and Javascript. It was so easy to me that I felt like I was fooling people taking their money. I also hustled part time as a telemarketer, phone support, salesman, working at caf├ęs and clothing stores.

  • 2009

    High Scool

    Started the Technology Programme at Fyrisskolan, Uppsala. I will never forget when my first year high school math teacher saying "You excel at the difficult thinking problems, but you rush by the simple ones, so I have to give you an E-". It scared and motivated me. If I wanted to work with programming, math was important. After almost failing, I didn't go to school for a month. Instead, I sat at home and did every single problem from that book and the next course book. I got a B on my next final.

  • 2009-2011

    iPhone, Software, Tools & 1st Real Programming Class

    I got the iPhone 4 in 2010 & learned how Jailbreaking and terminal worked. Started a club movement with a few friends, Club Strobe, aimed for 16-18 year olds. During High School, we took courses that taught us AutoDesk AutoCAD and SolidEdge 3D, Adobe Creative Suite, Google Sketchup, Adobe Flash - a bunch multimedia stuff. But the moment finally came, it was time for "Programming 1" and JavaScript! We got to create webapps. Calculator, Tic Tac Toe, Memory and a project. I extended my curriculum to take "Programming 2" individually - Object Oriented Programming with C#.

  • 2012

    University & First Company

    During the summer of 2012 I decided to create my own company, focusing on selling fashionable accessories - Fett . I got admitted to the Bachelor's programme in Computer Science at Uppsala University, but not to KTH I really wanted. I had to put my startup Fett on hold so I could prioritize my studies. I was offered a year long full-time role at IBM and Ericsson, but I wanted to focus on my studies, so I turned them down. I learned Java and C++ at Uppsala University. I Decided that I wanted to move to Stockholm & applied a second time to KTH and got admitted. Goal!

  • 2013

    Microsoft Internship

    I left Windows behind me and got a MacBook Air before starting my second year of University at KTH. Ironically, during the fall of 2013 I landed my first industry job, and it was with the company I had always dreamt of working for, namely Microsoft. Bill Gates was my hero and main idol. I got to do Windows/Windows Phone 8 & 8.1 application development with Visual Studio & C#. I further developed a C# application that generated social media applications, using RSS feeds, Twitter, Youtube, Instagram and other feeds.

  • 2014-2015

    Startup, Side projects & Machine Learning

    Developed an Android application to steer UHQ cameras installed in a visualization studio for Software Engineering class at KTH with some friends. Created the cross platform game Angry Clouds with a good friend of mine. Learned Angular/Node/MongoDB for a class project/sideproject to remove the video part from Youtube and only focus on music. Wrote a shell in C. Wrote my Bachelor's thesis in the field of Machine Learning, and fell in love with stats again. Learned & implemented Linear Regression, Naive Bayes, Decision Tree, Random Forest & Artificial Neural Network classifiers for Stock Prediciton.

  • 2015

    Summer

    Moved to London to intern for EF Education First without knowing what I was going to work on. Learned Git, iOS Development in Objective-C and Swift & pushed my first production ready feature after 2 weeks. Learned all about a multi-cultural workplace, Agile Development/SCRUM and how to develop with scale & production in mind. Worked on EF Class with the CTX Team for the rest of the summer.

  • 2015 - 2016

    From fall to summer

    Moved to San Luis Obispo, California for a year on exchange and specialized in Machine Learning. Had the best year of my life at CalPoly SLO.

  • 2016

    Summer

    Moved to San Fransisco & Interned with the Siri Platform Team in Cupertino, California. Got to learn how to do Machine Learning in production, how Intelligent Personal Assistants work & how to scale it. Worked on three intern projects with iOS, watchOS, backend & web development in C/C++/Objective-C/Swift, Python/SQL/HTML/CSS. Worked with high priority problems with the core watchOS team and worked on test automation with the Siri Media Domains team. Full Stack Deluxe.

  • 2017

    Fall to Summer

    Came back home, pitched a project idea to Spotify & they liked it. Worked as a research intern in the area of Machine Learning at Spotify. I researched & implemented deep neural networks in TensorFlow for the task of music recommendation, exploration and search. Accidentally built a system that detects fraud in the process.

  • 2017

    Summer

    Defended my Master's thesis Deep Neural Networks for Context Aware Personalized Music Recommendation I had been working on at Spotify.

  • From script kiddie to a Master

    Goal!

Education

education & degrees

Work Experience

Places I've worked at

Work Experience

Lifesum

2018-
Platform Engineer

Serving as Platform Engineer at Lifesum. Fullstack developing featuring backend, data and machine learning engineering.

Learn more

Tictail

2017
Machine Learning Engineer

Serving as Head Machine Learning Engineer at Tictail. Fullstack developing featuring backend, data and machine learning engineering.

Learn more

Spotify

2017
Machine Learning

I worked at Spotify with the Search team in Stockholm, Sweden for six months, developing & researching neural networks for the task of recommendation and search.

Learn more

Apple

2016
Software Engineering

In Cupertino, California, I got to work at Apple for the Siri Platform Team for over three months, developing several platform tools for managing and analysing Siri data.

Learn more

Microsoft

2013
Software Engineering

At Microsoft, I got the chance to learn about and develop mobile Windows/Windows Phone 8/8.1 Applications and learn a lot about automation and data generation.

Learn more

Education First

2015
Software Engineering

After teaching for over a year at university, I got the chance to develop iOS tools to do so, at scale, with the CTX team at EF Education First in London.

Learn more

KTH

2014-
Ambassador & TA

Acting as ambassador & blogger for my university for years. Been fortunate enough to be a TA for fundamental CS and business classes.

Learn more

Project Portfolio

A selected portfolio of my coding projects
×

Deep Neural Networks for Context Aware Music Recommendation

Deep Neural Networks for Context Aware Music Recommendation

Master's Thesis /
Spotify Machine Learning Research Internship Project

Abstract

Information Filtering (IF) and Recommender Systems (RS) have been used and has been implemented in various ways from various entities since the dawn of the Internet, and state-of-the-art approaches rely on Machine Learning (ML) and Deep Learning (DL) in order to create accurate and personalized recommendations for users in a given context. These models require big amounts of data with a variety of features such as time, location and user data in order to find correlations and patterns that other classical models such as matrix factorization and collaborative filtering cannot. This thesis researches, implements and compares a variety of models with the primary focus of Machine Learning and Deep Learning for the task of music recommendation and do so successfully by representing the task of recommendation as a multi-class extreme classification task with 100 000 distinct labels. By comparing fourteen different experiments, all implemented models successfully learn features such as time, location, user features and previous listening history in order to create context-aware personalized music predictions, and solves the cold start problem by using user demographic information, where the best model being capable of capturing the intended label in its top 100 list of recommended items for more than 1/3 of the unseen data in an offline evaluation, when evaluating on randomly selected examples from the unseen following week.

Project Example

Netflix and Predict

Netflix & Predict

AI Class Project at CalPoly SLO

Together with two other classmates, we scraped data from IMDB and used a Naive Bayes classifier in order to predict future ratings, given the show rating, and the previous episode-specific ratings. Using a Naive Bayes model to predict future episode specific as well as overall rating for chosen TV shows such as The Walking Dead, Big Bang Theory, South Park, Heroes Reborn, South Park, Modern Family, Arrow.
Accuracy differs, with TWD and BBT with the best scores.

Github Repo

Project Example

Stock Market Prediction using Social Media Analysis

Stock Market Prediction using Social Media Analysis

Bachelor's Thesis

Implemented classical machine learning methods for predicting stock prices with Twitter & Yahoo! Stocks data.

Abstract

Stock Forecasting is commonly used in different forms everyday in order to predict stock prices. Sentiment Analysis (SA), Machine Learning (ML) and Data Mining (DM) are techniques that have recently become popular in analyzing public emotion in order to predict future stock prices. The algorithms need data in big sets to detect patterns, and the data has been collected through a live stream for the tweet data, together with web scraping for the stock data. This study examined how three organization's stocks correlate with the public opinion of them on the social networking platform, Twitter. Implementing various machine learning and classification models such as the Artificial Neural Network we successfully implemented a company-specific model capable of predicting stock price movement with 80% accuracy.

Publication Link

Project Example

Angry Clouds

Angry Clouds - a game about balance! Keep the rain away!

Created a fun & simple cross platform game for iOS and Android with one of my good friends Ivan. Angry Clouds reached #31 in the Sweden App Store under Platform games.

App Store Download

Google Play Download

Project Example

GLIP

Created a project acting as frontend / backend developer during a Hackathon together with four other team members at a Spotify Hackathon in the Stockholm HQ. Together we created an application for generating a GIF video for a given song on Spotify. We used a lyrics API to request gifs matching a word, and do our best to sync the gifs and lyrics to the song. After the hackathon, we found out that the lyrics are in sync with the YouTube version of the song, and will not sync with all album track versions. It was a great accomplishment for a project I had been thinking of doing for a long time.

Check GLIP out

Project Example

Bioinformatics & Sea Stars

Bioinformatics

In four months time, I took Bioinformatics Algorithms at CalPoly SLO, which turned out to be a project class. Together with two other computer science students, we built a team and had another team of biochem majors, assigning us real tasks, simulating a data science/lab situation. We did DNA Analysis on CG-Content, Gene Density, Local & Global Alignment with BLAST and FASTA, and learned about sequence alignment of proteins using BLOSUM. We did gene prediction and learned about hidden markov models to do so. I had a great time throughout the course of the project, thanks to some great teammates. This course really opened up my eye to bioinformatics and is something I have great passion for after taking this class.

Sea Stars

For the graduate HCI class at CalPoly, me and three other students worked on an iPad app for assisting the CalPoly Marine Biology Center collect & analyze data under water. We even got to see them do it in action, and had a streamline feedback loop regarding the ease of use (easy UI, size of elements, minor text input), to finally create a report consisting of the state & health of the local Sea Stars. Awesome stuff.

Bioinformatics Github Repo

Sea Stars Github Repo

Design Portfolio

portfolio of some creative design work

Design Portfolio

×

Project Example

AutoCAD

Previous CAD 2D / 3D Projects

As required by the technology programme I was enrolled in High School, we had three classes where we learned how to do complex AutoCAD models, both in 2D and 3D. Furthermore, we took a basic course in Mechanical Design and actually got to create one of these designs with our bare hands, and program the moves for a machine to do so. Sick!

Project Example

Fett

Artwork & Merch Display from my solo accessory startup

I started my own Merch fashion line called Fett. I basically put beads on $2 alibaba iPhone headphones and sold them for around 10 times more. I put my entire family to work and sold around 10.000 of them in total, they were pretty popular. Good ROI and a fun summer 2012 sidehustle, but not sustainable. I gave up Fett when I started university, Fall 2012.

Instagram

Project Example

Coverart Work

Some previous original coverart work

Since I've been using Photoshop since 10 years of age, I've done everything myself when I had to. It's just a tool, afterall. I did some fan coverart, but also did some professional freelancer side hustle work.

Project Example

Websites

Some previous website designs

Some past website designs I've designed and created.

×

Project Example

Deep Neural Networks for Context Aware Music Recommendation

Researched & implemented embedding/feed forward neural networks.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nam sapien risus, blandit at fringilla ac, varius sed dolor. Donec augue lacus, vulputate sed consectetur facilisis, interdum pharetra ligula. Nulla suscipit erat nibh, ut porttitor nisl dapibus eu.

Phasellus porta eros vel lacus euismod consequat. Phasellus eleifend, nibh non feugiat hendrerit, lacus enim adipiscing eros, a pharetra erat neque eget est. Quisque vitae aliquet urna. Integer suscipit lectus eu enim porttitor fringilla. Ut blandit, urna in auctor euismod, arcu eros pharetra metus, a porta purus libero a nibh.

Nam eget urna pellentesque nisl ultrices dapibus. Mauris accumsan vehicula nisl, sed tempus mauris facilisis eu. Donec a iaculis nisi, quis malesuada justo. Pellentesque ut enim sit amet ipsum dignissim egestas. Morbi tincidunt rhoncus urna eget placerat.

Visit Website

Project Example

Deep Neural Networks for Context Aware Music Recommendation

Researched & implemented embedding/feed forward neural networks.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nam sapien risus, blandit at fringilla ac, varius sed dolor. Donec augue lacus, vulputate sed consectetur facilisis, interdum pharetra ligula. Nulla suscipit erat nibh, ut porttitor nisl dapibus eu.

Phasellus porta eros vel lacus euismod consequat. Phasellus eleifend, nibh non feugiat hendrerit, lacus enim adipiscing eros, a pharetra erat neque eget est. Quisque vitae aliquet urna. Integer suscipit lectus eu enim porttitor fringilla. Ut blandit, urna in auctor euismod, arcu eros pharetra metus, a porta purus libero a nibh.

Nam eget urna pellentesque nisl ultrices dapibus. Mauris accumsan vehicula nisl, sed tempus mauris facilisis eu. Donec a iaculis nisi, quis malesuada justo. Pellentesque ut enim sit amet ipsum dignissim egestas. Morbi tincidunt rhoncus urna eget placerat.

Visit Website

Project Example

Deep Neural Networks for Context Aware Music Recommendation

Researched & implemented embedding/feed forward neural networks.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nam sapien risus, blandit at fringilla ac, varius sed dolor. Donec augue lacus, vulputate sed consectetur facilisis, interdum pharetra ligula. Nulla suscipit erat nibh, ut porttitor nisl dapibus eu.

Phasellus porta eros vel lacus euismod consequat. Phasellus eleifend, nibh non feugiat hendrerit, lacus enim adipiscing eros, a pharetra erat neque eget est. Quisque vitae aliquet urna. Integer suscipit lectus eu enim porttitor fringilla. Ut blandit, urna in auctor euismod, arcu eros pharetra metus, a porta purus libero a nibh.

Nam eget urna pellentesque nisl ultrices dapibus. Mauris accumsan vehicula nisl, sed tempus mauris facilisis eu. Donec a iaculis nisi, quis malesuada justo. Pellentesque ut enim sit amet ipsum dignissim egestas. Morbi tincidunt rhoncus urna eget placerat.

Visit Website

Project Example

Deep Neural Networks for Context Aware Music Recommendation

Researched & implemented embedding/feed forward neural networks.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nam sapien risus, blandit at fringilla ac, varius sed dolor. Donec augue lacus, vulputate sed consectetur facilisis, interdum pharetra ligula. Nulla suscipit erat nibh, ut porttitor nisl dapibus eu.

Phasellus porta eros vel lacus euismod consequat. Phasellus eleifend, nibh non feugiat hendrerit, lacus enim adipiscing eros, a pharetra erat neque eget est. Quisque vitae aliquet urna. Integer suscipit lectus eu enim porttitor fringilla. Ut blandit, urna in auctor euismod, arcu eros pharetra metus, a porta purus libero a nibh.

Nam eget urna pellentesque nisl ultrices dapibus. Mauris accumsan vehicula nisl, sed tempus mauris facilisis eu. Donec a iaculis nisi, quis malesuada justo. Pellentesque ut enim sit amet ipsum dignissim egestas. Morbi tincidunt rhoncus urna eget placerat.

Visit Website

Medium Blog

@oktaybahceci

KTH Ambassador Blog

kth.se/blogs/oktayb