Peut être disponible
(Mis à jour 2022-08-22)Machine Learning Engineer
Taastrup, Danmark
Natif Danish, Courant English
- Data Science
- Natural Language Processing
- Machine Learning
Compétences (27)
SUPERVISED LEARNING
Python
Artificial Intelligence
NEURAL NETWORKS
Deep Learning
NLP
Machine Learning
PANDAS
data
KERAS
COMPUTER VISION
Natural Language Processing
TENSORFLOW
AI
PATTERN RECOGNITION
SUPPORT VECTOR MACHINE
DATA VISUALIZATION
DATA SCIENCE
Analysis
SCRAPING
Forretningsforståelse
MongoDB
TEXT-TO-SPEECH
Experience
Git
Elasticsearch
SQL
Résumé
I am a machine learning engineer with experience in Python, Tensorflow and Keras. I have worked with NLP, classification and regression professionally, and build AI software for multiple different companies. Here I achieved to beat state-of-the-art Danish text analysis, in order to help companies with GDPR, build state-of-the-art Danish property valuation AI, and I've created novel algorithms for advanced game AI. Lately I took on an international adventure to Barcelona, working on Apple's digital assistant Siri, mainly for the Danish users. My focus has always been on deep learning, and its different branches. I have hands-on experience with everything from basic neural networks, to convolutional nets and LSTMs. Coming from a university with focus on problem solving and working in teams, I can bring a real team member to your company, along with a deep passion for machine learning, and an eagerness to learn everything within the field. I also have 4 years of experience with creating games, where I established and led a team of game developers. I had multiple games reach top 5 in worldwide game jam contests, one game getting featured on Newgrounds.com and one game nominated for game of the year on Gamejolt.com. I'm used to work in teams, whether it's in a physical room or across the globe through Skype.
Expérience professionnelle
2019-07 - 2020-01
I was headhunted to Barcelona, to help make Siri the best digital assistant worldwide, and specifically to improve user experience for the Danish locale. The project ranged from automatic speech recognition, building natural language processing components for intent prediction, and finally user feedback through text-to-speech. I was part of a smaller Danish team, where I helped tackle complex NLP and ML problems using large amount of data, as well as analysing and making data driven decisions to improve Danish user experience.
Methods - NLP, Text-to-Speech, Speech Recognition, Big Data. (Not allowed to share more specifically)
Technologies - Not allowed to share
2018-10 - 2019-07
I've worked on a classifier for Danish government agencies, which can automatically journalize emails and documents, received by citizens. Citizens send tons of information, in many varied and unstructured forms to the government. Each of these must be recorded, and assigned the proper case number.
Methods - NLP, RPA, NER, PoS, Deep Learning, Computer Vision, Convolutional Neural Networks, Transfer Learning, Word Embedding, Web Scraping, Feature Engineering.
Technologies - Python, Keras, Tensorflow, Gensim, sklearn, Polyglot, Github, Docker
2019-03 - 2019-07
Helped establish the core requirements for an advanced recommender system. I was brought in, to give my advice and ideas as to how the AI system should be approached, and what data was needed for it to work. I further investigated the company's options for using different tools, and to which extend these could satisfy the company's needs.
Methods - Recommender Systems, Time-Series, Sequential Data, (Hierarchical) Recurrent Neural Network, Collaborative and Content-based filtering.
Technologies - Amazon Personalize
2018-10 - 2019-07
I've created a text-analysis tool for one of the biggest financial institutions in Denmark, to help them discover private information in their data, so they were able to comply with GDPR. The biggest challenge in this was project was the lack of Danish datasets, which forced me to find other solutions. My AI ended up beating state-of-the-art, a NER created by Copenhagen University.
Methods - NLP, NER, PoS, Deep Learning, Word Embedding, SVM, Naïve Bayes, Regex, Web Scraping, Feature Engineering.
Technologies - Python, Keras, Tensorflow, Gensim, sklearn, Polyglot, Github, Java, ElasticSearch.
2018-10 - 2019-07
I was asked to help develop an automatic property valuation model. In the first month,
my AI was better than their own model, which they've worked on for 2 years. After a
couple of months, I've built one of the best, if not the best automatic property valuation
model in Denmark, to help them make the best mobile Real Estate App for the Danish
market.
Methods - Deep Learning, Feature Engineering, Big Data, Data Visualization.
Technologies - Python, Keras, Tensorflow, Pymongo, Pandas, Seaborn, sklearn,
MongoDB, Github, Docker.
Parcours scolaire
2016-01 - 2018-01
2013-01 - 2016-01