Daniel Licht
Data Scientist | Educator | Astrophysicist
Education
University of Colorado Boulder
Master of Science Astrophysical and Planetary Sciences 2006
GPA: 3.7
University of Rochester
Bachelor of Science Physics & Astronomy 2003
Minors: Linguistics, Mathematics
Cum Laude, Cumulative GPA: 3.7
Skills
Data Analysis: Python, Numpy, Pandas, SciPy, AWS, SQL, Image Processing, GeoPandas, BeautifulSoup, NoSQL, Apache PySpark
Machine Learning: Scikit-Learn, Clustering, Classification, Natural Language Processing, NLTK, Word2Vec, Regression, Boosted Trees, Deep Learning, Keras, TensorFlow, Semantic and Image Segmentation
Data Visualization: Matplotlib, D3.js, Seaborn
Teaching: Credentialed in Physics, Earth Sciences, Chemistry
Recent Data Projects
Geo-Spacial Analysis with DigitalGlobe

Trained an image segmentation U-net to map out flooded areas in a newly released Hurricane Harvey dataset from DigitalGlobe, in collaboration with DigitalGlobe data scientists.  Designed a semi-supervised process to generate pixel-level image labels from a course map of possible flooding locations .

Mapping of Character Interactions using NLP

Used Natural Language Processing techniques, including NLTK, parsing, word2vec, clustering, and sentiment analysis to generate a pipeline to identify primary characters in a novel or series of novels, characterize the characters and/or relationships between them, and map out the frequency of interaction and progression over time. Data was visualized using several interactive D3.js applications.

Classification of Gestures from SmartWatch Accelerometer Data

Used snippets of smartwatch accelerometer data to identify one of 20 gestures. This required generating all features from the accelerometer data. Achieved 98% recognition accuracy on the dataset using both individual and ensemble models, if permitted to train on some of a user's previous actions; and 88% accuracy for users the system has never seen before.

Experience
Metis
San Francisco, CA
Data Scientist
2017 to Current

Part of a competitive 12-week immersive Data Science program covering topics in statistics, machine learning, programming, communication, and design.

  • Primarily working in Python, completed several end-to-end data science projects involving a suite of Machine Learning applications, including Natural Language Processing, Deep Learning, and interactive D3.js visualizations.

Fremont Unified School District
Fremont, CA
AP Physics and Engineering Teacher
2012 to 2017
  • Taught data analysis skills to science students, including curve fitting, linearization, and uncertainty estimation.
  • Spearheaded the creation of a new AP Physics C course never before offered at the school.
  • Became a Project Lead The Way (PLTW) certified engineering and IED STEM education instructor in 2014.
  • Led district-wide adoption of new technologies for the classroom, by being the 1st teacher in the district to adopt and assist with the wider deployment of an online assessment platform.
  • Coached competitive Science Bowl and Science Olympiad teams (ranked 2nd and 7th in Alameda county, Spring 2017)
UC Berkeley - Academic Talent Development Program (ATDP)
Berkeley, CA
Instructor - Secondary Division (summers)
2015 to 2017

  • Taught highly accelerated college-level courses for academically gifted high school students (Astrophysics & ElectroMagnetism).

Pittsburg Unified School District
Pittsburg, CA
High School Physics Teacher
2011 to 2012

  • Continuously developed new curriculum in collaboration with other teachers, including department-wide common assessments data collection.
  • Worked with advanced level ELL students, at-risk, and economically disadvantaged youth.

Caltrol Inc.
Pleasanton, CA
Systems Engineer II: Industrial Field Controls
2006 to 2010

  • Created graphical UI and programmed underlying system control logic for DeltaV control systems.

  • Trained plant engineers and operators in the use of the programing and control software.