Experience

Autobrains Technologies

• Israel
• Software Development
• 100 - 200 Employee

• Deep Learning Engineer

  • Jan 2022 - Present

  • Training novel end-to-end driving policy models and specific driving functions using supervised deep learning and deep reinforcement learning approaches. • Generating, collecting, augmenting, and preprocessing image and kinematic training data. • Reward engineering, model hyper-parameter tuning, ablation studies, real2sim studies, and algorithm optimization. • Designing, building, and automating data extraction/generation pipeline on AWS • Setting up local Linux/Windows machines and AMIs on AWS for different training environments. • Building simulation environment in ROS, RViz, Pygame, and CARLA for data generation, model training, and model evaluation. • Developing KPIs, benchmarking, versioning, and analysis of trained models. • Image processing for visualizing model outputs and creating birds eye view and image plane animations for presenting results and improving model explainability. • Writing and contributing to 15+ filed patents related to team’s technological innovations. • Python, PyTorch, OpenAI Gym, OpenCV, NumPy, Pandas, Matplotlib, Scikit-learn, CARLA, ROS/ RViz, AWS, git, LaTeX Show less



Natural Intelligence

• Israel
• Software Development
• 400 - 500 Employee

• NLP Data Science Intern

  • Nov 2021 - Dec 2021

  • Implemented deep learning NLP models to generate hundreds of high quality content articles related to services on top10.com platform using GPT-family models. • Work included fine-tuning models on AWS with parallel GPUs, hyperparameter optimization, and scraping relevant textual data from the web for fine-tuning. • Developed an evaluation method to flag problematic generated texts drawing from BLEURT text generation evaluation model • Python, Huggingface, AWS, NLTK, Pandas, SQL, DeepSpeed, BeautifulSoup Show less



Reichman University

• Higher Education
• 700 & Above Employee

• Data Scientist

  • Mar 2021 - Oct 2021

  • Developed and analyzed linear models and applied several clustering algorithms to better understand the relationship between material stock distributions and social, economic and environmental factors at the NUTS3 level across Europe. • Collected, cleaned, and preprocessed data before modeling. Produced visualizations of results. • Python, Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn, QGIS • Developed and analyzed linear models and applied several clustering algorithms to better understand the relationship between material stock distributions and social, economic and environmental factors at the NUTS3 level across Europe. • Collected, cleaned, and preprocessed data before modeling. Produced visualizations of results. • Python, Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn, QGIS



Self-employed

• Budapest, Hungary

• Musician

  • Oct 2013 - Oct 2020

  • Travelled across Europe, USA, and Asia touring with several Hungarian folk ensembles based out of Budapest, Hungary • Travelled across Europe, USA, and Asia touring with several Hungarian folk ensembles based out of Budapest, Hungary



University of Maryland

• United States
• Higher Education
• 700 & Above Employee

• Teaching Assistant (Electronics & Instrumentation II)

  • Sep 2012 - May 2013

  • Prepared and delivered lectures and lab sessions to sophomore and junior electrical engineering students• Supported and worked with students during lab sessions and office hours

• Graduate Researcher (MEMS and Microfluidics Laboratory)

  • Jan 2011 - Feb 2013

  • Modeled and fabricated piezoelectric micro-transformers• Designed successful photolithographic and etching processes for bulk PZT