## Bio Technologist. Interested in machine learning, A.I. and tech in general. I work as a scrum product owner & technical project manager and do various projects in my free time. Always looking to learn new things. Reach me at [Linkedin](https://dk.linkedin.com/in/asbjoern-andersen-b10a3a36) or directly at as.svng.an@gmail.com - I'm always open to discuss ideas and concepts! I always tinker with various projects and ideas. Both digitally and offline. ## Digital **Machine Learning by Stanford University ([via Coursera](https://www.coursera.org/learn/machine-learning)** - *2017 march - present* Currently enrolled into the 11 week online Coursera course "Machine Learning" where professor Andrew Ng teaches all things machine learning. It is a great course that preps one to many different types of machine learning algorithms, the underlying theory as well as concrete practical hands-on practice. **Retell** - *app, cogntive APIs, 2017 march - present* Latest app that I've been working on is Retell. It's a service for those who commute and wish to listen to interesting articles. Basically a Spotify meets intriguing articles with daily digests on the go. It uses various text extraction APIs to extract text from urls, Amazon Polly API to generate audio and a Rails backend to concatenate audio files. An example of the audio generated can be heard [here](https://voice-io.s3-eu-west-1.amazonaws.com/47_6eb135389a.mp3). **Cryptocoin investing and ML** - *machine learning, data mining, cryptocurrencies, 2016 oct - present* To experiment with machine learning and predictions I decided to try my efforts on the field of cryptocoins. Aim is to investigate whether public opinion can be used to predict the value of currency. I thus began to scrape prices for some 500 crytocoins as well as all tweets (alongside sentiment analysis) related to each currency. As of march 2017 the database consists of 3 mio. tweets og 2.5 mio price points. Based on the data different types of feature sets are used to see if predictions are possible. Built with Ruby on Rails, Google Predictions API and Twitter API. **[Madmarkedet.dk](http://madmarkedet.dk)** - *founder, startup, entrepreneurship, 2013 - 2016* In 2013 I co-founded a food startup that set out to change the way consumers bought food from producers. A type of "Etsy for foods" giving the producers a set of tools and freedown to set their own prices. At its peak we had 160 products and 16 producers. After 3 years as a great passion project(having worked on it nights and weekends) we decided to pause it in june 2016 to pursue other projects. **Holy carp** - *app, cogntive APIs, data mining, 2017 jan* As a fun (and silly) experiment of what the latest cognitive computing APIs could be used: - I scraped top 1000 jokes from reddit.com/r/jokes & categorize them into keywords/entities using Aylien.com entity analysis API - Created a Messenger chatbot where users could upload an image - Used Microsoft Cogntive APIs to figure out what was on the image - Return a joke that matched the content of the joke based on a fuzzy string similarity algoritm, e.g. 'outdoor' is close to 'door' - For an example of Holy Carp in action: [link](https://s3-eu-west-1.amazonaws.com/asbjoernandersen.com/holy+carp.png) **[Notedown.io](http://notedown.io/)** - *tool, PM, productivity, 2016 june* As a PM I grew tired of complex task managers and instead chose to go way back to the basics: a minimal notes and task manager using markdown. Built with Javascript. **[Send.digital](http://send.digital/)** - *tool, productivity, 2016 april* A simple means of sharing files through Ctrl + V or simply dragging a file on the site. Built with Javascript, Ruby on Rails and the Amazon SDKs. **On the topic of A.I, NLPs and cognitive computing** - After having read Ray Kurzweil "How to Create a Mind: The Secret of Human Thought Revealed" I am deeply fascinated by the idea of creating a digital self. I don't have an exact means to get there but I am fascinated by the idea of an AI that has a real mind. Could be an assistant that inherits a persons own persona to whom one may outsource the trivial tasks. E.g. by training it with your every move and decision. AI has to learn what it means to be you. - The subject of the ethics behind the rapid development of A.I. E.g. how do we make sure that humanity still thrives in a future where many jobs will get automated. - Knowledge consumption and NLP. Training an A.I. algorithm with rich texts in order to make sense of it. Having come across this [repository](https://medium.com/startup-grind/fueling-the-ai-gold-rush-7ae438505bc2#.vpa6fw6uf) and Facebook open sourcing their [fasttext vectors](https://github.com/facebookresearch/fastText/blob/master/pretrained-vectors.md) perhaps one should have a go at it. ## Offline **Range Rover Classic** - I'm a big fan of old cars. I have a '75 Renault 4 and last year I bought a 92' Range Rover Classic which I've been spent a fair amount of time renovating. I love the outdoors and the RRC is used to take us to great places. I am a member of the Danish Land Rover community (DLRK) and I post photos of the travels on [Instagram](https://www.instagram.com/exploringinmyrangerover/) **Drone building** - I've always had a thing for building things and remote controlled cars. Only natural that I took an interest in drone building and flying. Instead of buying a ready-made drone I went ahead and got myself a "kit" with all the individual components. I then spend a good month on soldering things together and calibrating firmware/sensors etc. before I got to actually flying it. When everything was put together: [video](https://www.dropbox.com/sc/k0v5xvdda77amkp/AABVVo8sFX42_J54raNqDmZBa) and this is one of my first flights: [video](https://www.dropbox.com/sc/ihfwk9w84svcz3d/AAAb5BzM9vlh3Lr96pB0IIiea) **Electronics building** - I've been experimenting with combining a chaotic double pendullum inside a box to generate power through a set of gears where dual rotations produce a rotation around only one axis. I got a basic prototype up and running and need to purchase a proper motor to see if the produced power has a feasible magnitude. Point would be have a power generating device to bring on hiking treks or similar.