Alexander Prochnow alexprochnow132@gmail.com E DUCATION University of Osnabrück (Bachelor of Cognitive Science) 2018 – September 2021 Gymnasium Raabeschule, Braunschweig (GPA 3.5/4, best of my year) 2010 – 2018 R ELEVANT E XPERIENCES Here I listed some of my projects and practical experiences. I implemented and applied most of the algorithms from scratch, as part of weekly homework throughout my studies. During classes and group work I often presented my ML solutions and can explain solution approaches, concepts and algorithms on a technical as well as on a conceptual level intuitively. Deep Reinforcement Learning project: As final project for my TensorFlow course we implemented the Deep Reinforcement Learning algorithm a2c (synchronous advantage actor-critic) and trained it on the Lunar Lander Continuous environment from OpenAI Gym. We multi-threaded the training by creating multiple agents and synchronizing their training every few timesteps to boost training speed. We were able to solve the game by OpenAI’s definition after around 2500 training episodes. (Link to the project: https://github.com/jootten/A2C_Lunar_Lander). Deep Learning: Practical experience with TensorFlow/Keras and PyTorch, working on e.g. adversarial training, attention, GANs, ResNet, BERT, Transformer, LSTMs, CNNs etc. with strong understanding of the fundamentals. Computer Vision: for example image enhancement, morphological operators, image labeling, segmentation, template-matching algorithms (using Numpy, OpenCV, scikit-learn/image), compression, motion analysis (e.g. Horn-Schunk, Lucas-Kanade), pattern/object recognition (e.g. SIFT, SURF, Viola-Jones, Wire-frame model), CNNs etc. Machine Learning: Data Mining (e.g. outlier detection, missing values estimation), clustering algorithms(e.g. EM- algorithm, k-means), dimension reduction (e.g. PCA, projections, multidimensional scaling), local methods (self-organizing maps, radial basis functions), classification (e.g. Support vector machine, random forests, ID3), reinforcement learning (e.g. Q-Learning, Markov Decision Processes, Monte-Carlo), combinatorial optimization (e.g. linear programming, flow networks, genetic algorithms, constraint programming), search algorithms, planning (e.g. Situation calculus, STRIPS) etc. Natural Language Processing: Methods for word embeddings (e.g. GloVe, Word2Vec, CBoW etc.), applying common neural network architectures to NLP problems using PyTorch and TensorFlow, e.g. Transformer, Seq2Seq, LSTMs, GRUs, bidirectional RNNs etc. solving problems in the fields of Natural Language Generation, Sentiment Analysis, Information Retrieval, Summarization, Recommender Systems, Dialog Systems and Machine Translation. Data Science and Data Analysis: Frequentist and Bayesian data analysis, Hypothesis testing, Bayesian regression models. As a semester-long project in experimental design and data analysis, we replicated an experiment from a published experimental psychology paper with N=72 participants. Here I learned how to clean and analyze self-collected data in order to dispute the validity of the original paper in a quantifiable manner. The focus here was on Open Science and the reproducible portrayal of analysis results. (Link to the project: https://github.com/Group1XPLab2020/Size_and_Space_exp1_magpie_Replication). Tools I am familiar with for statistical data analysis: R (Packages: tidyverse for data cleaning, ggplot for visualization, rstan for probabilistic modeling, brms for Bayesian generalized regression models) as well as Python (pandas for data cleaning/analysis, matplotlib/seaborn for visualization) Throughout my studies I analyzed data and trained models in many different fields, including various industrial data sets, EEG and eye tracking data, medical imagery, sensor data (e.g. from various agricultural soil sensors), satellite images, experimental psychology and survey data etc. I am familiar with cloud computing using custom hardware via the universities computer grid and CUDA O THER E XPERIENCES AND I NTERNSHIPS Globalink Research Internship at Concordia University, Montreal, Canada (May 2021 - July 2021): The Canadian national organization Mitacs offers sponsored three month internships at Canadian research institutes and universities. Due to travel restrictions, my internship is taking place virtually. As part of the internship I am writing my bachelor’ s thesis in automated testing for Deep Learning frameworks. For this I am analyzing current software testing practices of TensorFlow, PyTorch, Keras and Theano and developing an automated tool that should aid the framework developers with maintaining and improving their testing and quality assurance. Working student in software development at InfinityGate (September 2020 – ongoing): The startup makes software for designers, constructors and architects that allows them to view and change their models in photorealistic virtual reality for rapid, cost-effective prototyping. I am reworking the online subsystem and streamlining the online features, which allow multiple users to collaboratively view models. I am also working on a novel mesh modifier using experimental technology from the Unreal Engine. For this I am in regular contact with the Unreal developers as well as with clients for feedback during development. (C++). Paid internship at Design Ready Controls, Minnesota (August 2018): The company engineers and manufactures system controls. The internship taught me CAD, as well as practical engineering skills. Student Engineering Academy (SIA) (February 2016 – December 2016): Statewide project for students interested in science and technology. Main constituent were the visits to companies and institutes every Friday, including Siemens and the German Aerospace Center. Information Technology Project (ITech3) (2017): Fellow students and I had three months to plan and build a science project for a local student fair. At the fair we presented our LED-panel version of the popular game Snake and explained the functionality to visitors. Here I learned how to program Arduino microcontrollers and access sensor data. Internship at CodeFrog IT consulting and service provider (January 2016): Working there taught me my first programming language, Java, which I became proficient in during the three-week internship. It also gave me insights into how programming projects are executed in bigger teams. Mentor for nine Cognitive Science freshmen (winter semester 2020): I was able to help a group of freshmen during their first semester, who would have otherwise had a difficult start into an online semester due to Covid. I supported them during their studies and organized social events to make it easier for them to get in contact with other freshmen, since there were no in-person lectures. Leader of the stage management group for our high school theatre; Student body treasurer; High school graduation committee. G ENERAL S KILLS Languages: German (native) English (native) Latin (educ. qualification) EDV: Programming/scripting languages: Python, R, Java, C, C++, C# Swift, SQL, Prolog, JavaScript, Bash scripting Software: Git, Perforce , Jupyter Notebooks, software testing/build automation (e.g. Maven), Unity game engine (2D/3D/VR), Unreal Engine (2D/3D/VR), Arduino programming, Adobe Photoshop, Adobe Premiere Pro, MS Word, PowerPoint and Excel, MacOS, basic familiarity with Linux I NTERESTS ( OUTSIDE OF AI) Personal coding projects, for example 2D/3D/VR games made with the Unity game engine. Working on start-up ideas with friends. Computer gaming: Currently StarCraft2, World of Warcraft. Tabletop games: Pen-and-paper role-playing games with friends, miniature wargaming, Magic: The Gathering. Personal fitness; parkour and freerunning; traveling; cooking Osnabrueck, 6/14/2021 Alexander Prochnow