Note for my English readers: This post is available only in German, since it describes my experiences specifically with job searching in Germany, and under German bureaucracy. I invite you to read it through a translation service of your trust,1 but please note that it’ll probably not be helpful or particularly insightful to you. Enjoy reading! So kriegst du einen Job als Software-Entwickler*in! Okay, stopp. Ich wünschte, ich könnte dir mehr bieten als diesen Clickbait-Titel, und es gäbe das Zauberrezept, und hurra, wir sind fertig! Leider ist es nicht ganz so einfach. Das Jahr neigt sich allmählich dem Ende zu, und ich selbst habe noch keinen Job gefunden.2 Allerdings kann ich dir hier ein paar Tipps dazu geben, was mir geholfen hat, zumindest entsprechendes Selbstwertgefühl (wieder) zu entwickeln, meinen Lebenslauf und mein Anschreiben ein bisschen zu optimieren, und was ich erfolglos probiert habe. Mit etwas Glück hilft dir, liebes Lesy3, mein Post damit weiter, deine eigene Suche ein bisschen erfolgreicher zu machen. ...
Icons Of Tech
Intro Women in Tech Awards and Recognition People Outside the Gender Binary
Prep a Course
You never know what resources you’ll need in the beginning, and that’s okay. Recently I did something that I’m a bit scared of, to be honest, and that is offer the local adult education centre my services as a teacher. I taught at my universities a little bit, but that was mostly tutoring, so I do think this is a bit different. When I started writing a concept for a Latin course I’m going to give, I had no idea what resources I would need. I first turned back to my own university materials: a comprehensive study grammar. Not much more, since I’m mostly using online dictionaries anyway (the really good dictionaries are either wildly expensive, or so incredibly extensive that you’d have to be very rich to buy them. Universities do that. I don’t. ...
Using UTF-16 Files in git – or How You Teach Linguists the Basics of Version Control
Intro This post is a bit of a story about how I came to know quite a lot about git, the version control system. The intended audience for this post is two-fold: Both technical folks like developers, who will likely be interested in the detailed section How To Actually Handle UTF-16 Files In git, and not quite as technical folks, who might be just generally interested in technology/IT topics, who might work as project managers but have no deep technical knowledge, or are just generally interested in whatever I write. ...
Recommendation Engines
This post serves as a memo towards building a content recommendation engine. –> There are 2 types of recommender systems: Collaborative Filtering — based on user rating and consumption to group similar users together, then to recommend products/services to users Content-Based Filtering — to make recommendations based on similar products/services according to their attributes. Calculate cosine similarity by using genre plot director main actors Keyword Extraction rake-nltk Limitations only languages supported in NLTK -> German? source code is quite old (last commit 8 months ago) SpaCy NER uses TF-IDF? (Source: Reddit) pre-trained word vectors -> quicker access to results but results might not be good enough (see here) Limitations min. 1k annotated utterances as training data (Source: Reddit): does this apply when we can simply use the pre-trained word vectors? Noun Phrase Extraction by Stanford CoreNLP uses parse tree to extract keywords (Source:Reddit) Recommendations Without User Preferences: A Natural Language Processing Approach Synopsis when information about user preferences is limited/not available approach based entirely on content of the recommended items comparing plot summaries of films (IMDB) topic signature genre similarity: not based on individual words in a plot summary but similarity to a whole set of film genres ...
Intro
Hello, welcome to my blog! It is currently under construction. I’m still having some issues with setting up and configuring Hugo, which I use for the HTML content creation. I hope to resolve the issues soon, so stay tuned for content! (Also I want to import my posts from my blog previously hosted at wordpress.com, which is a fun project but a bit of work in itself, so maybe soon there are going to be lots of interesting reads here!) ...
The hammer that turns every problem into a nail
Recently I asked a question on GitHub about some detail in the code of spaCy. Their founder of all the people, honnibal, called me condescending and presumptuous in return. I took offence at that at first, was irritated, but before I retorted with a snarky remark, I took a moment to breathe, and think about what he was actually saying. I realised that he was right. While I don’t consider myself to be condescending and/or presumptuous normally, my wording certainly had been: My question suggested I was telling the spaCy developers how to build an API for a complex python package – people who do this for a living, and whose code is recognised as one of the best NLP toolkits in the world. ...
Installing Packages in Jupyter-Notebook or Jupyter-Lab
Installing the packages in here Don’t run this code if you restart the kernel - run this once on setup of the virtual environment! Previously: pyenv virtualenv 3.8 lexKeySearch-3.8 echo "lexKeySearch-3.8" > .python-version # if you want to have spacy and the models in the venv, not only in the jupyter-lab, also execute these lines: pip install spacy python -m spacy download de_core_news_sm python -m spacy download de_core_news_md ipython kernel install --user --name='lexKeySearch-3.8' jupyter-lab Then, in the notebook, run the commands below in cells [1]-[8]. ...