Hershel Safer
Verified Expert in Engineering
Software Developer
Rehovot, Israel
Toptal member since January 20, 2021
Hershel has extensive experience developing quantitative models in various industries, including finance, biopharma, and communications. His focus is the extraction of data-based insights to inform decision making. For the past decade, Hershel has worked primarily with machine learning; he also has expertise in optimization (linear, nonlinear, combinatorial, and graph). Hershel has a Ph.D. from MIT in operations research and an undergraduate degree from Yale in applied mathematics and economics.
Portfolio
Experience
- Statistics - 20 years
- Algorithms - 20 years
- Combinatorial Optimization - 20 years
- Linear Programming - 20 years
- Machine Learning - 10 years
- Data Science - 10 years
- Python - 7 years
- Scikit-learn - 5 years
Availability
Preferred Environment
Jupyter Notebook, Scikit-learn, C++, R, Python, MacOS, Linux
The most amazing...
...tool I've developed was a trading model that combined machine learning and non-linear optimization, and automatically adapted to changing market conditions.
Work Experience
Founder | Machine Learning Consultant
Safer Analytics Ltd.
- Developed machine-learning models of credit risk and other issues that arise in business financing (merchant cash advances and invoice factoring).
- Built models for daily and high-frequency trading of equity and non-equity instruments, as well as tools for automatically creating models and ensuring that they work robustly.
- Presented a 2-day course in London for credit-risk professionals about using machine learning to increase their effectiveness.
- Mentored aspiring data scientists to enhance their ability to create tools for quantitative analysis.
- Consulted on techniques and tools of mathematical optimization to use for solving problems of product allocation and biomedical research.
Senior Quantitative Researcher
WorldQuant
- Developed and implemented computational models for statistical arbitrage on a variety of worldwide financial markets. Used a variety of techniques, including tools from machine learning and nonlinear optimization.
- Created a computerized system to automatically generate new trading models.
- Initiated an online system in the local office to facilitate information sharing among researchers.
Senior Research Fellow in Computer Science and Bioinformatics
Tel Aviv University
- Participated in research on computational methods for discovering cellular components that are involved in causing diseases. Used techniques from machine learning, combinatorial optimization, and graph theory.
- Wrote proposals for new research grants and progress reports for existing grants.
- Managed laboratory infrastructure that included various kinds of computing resources.
Head of Informatics
Dynamix Pharmaceuticals
- Participated in the design, implementation, and population of chemical-compound database for a new drug-discovery company.
- Managed the specifications and purchases of computers and storage.
- Oversaw the purchases and installations of software for computational chemistry.
Head of the Bioinformatics Core Facility
Weizmann Institute of Science
- Managed the core facility for data analysis in life-sciences research. Provided consulting, training, and computing resources for scientists at the Weizmann Institute and throughout Israel.
- Expanded the unit’s scope to serve a wider range of laboratories and added support for new research areas: next-generation DNA sequencing and protein structure.
- Managed the purchases and installations of computers and software for next-generation sequencing.
- Chaired the Fifth European Conference on Computational Biology (ECCB), the second-largest annual international conference in the field. The conference was a great success—400 delegates from 35 countries, and a surplus from our $290,000 budget.
Head of Informatics
Zetiq Technologies
- Managed and performed cheminformatics analyses for drug discovery.
- Designed and implemented software and database to gather, qualify, and store results from high-throughput screening assays.
- Developed analyses to identify interesting compounds for drug discovery.
Senior Scientist
Compugen
- Developed and implemented algorithms for designing genomics microarrays to profile the expression of splice variants.
- Led a group that developed and implemented algorithms for identifying proteins via mass spectrometry.
- Served as a project manager for the company’s largest customer.
- Wrote scientific articles and coordinated weekly seminar series.
Associate Director of Bioinformatics
Genome Therapeutics Corp.
- Developed a type of software to analyze genomic data for commercial clients and the Human Genome Project.
- Wrote the computational portions of two successful NIH grant applications.
- Helped build and led a computing group of ~30 people; managed the hiring and resolved personnel issues.
Experience
Weight of Evidence Target Encoding
https://github.com/hsafer/weight_of_evidenceWith unordered, non-numeric values, a feature like this cannot be used directly with most machine-learning methods. WOE addresses this challenge by converting each value, say "own with a mortgage," to the relative risk of an applicant in the group with that value paying off the loan. Specifically, WOE is the log odds of the categorical value being associated with a negative target value.
When the WOE of a categorical value is positive, the probability of the loan being paid in full is above average among all applicants, and vice versa when WOE is negative.
WOE has advantages over the one-hot encoding that is commonly used for categorical features. WOE avoids the proliferation of features when one-hot encoding is used with categorical features that take many values. WOE also does not lose information the way that one-hot encoding does when tree-based algorithms select a subset of features at each node.
Education
Ph.D. in Operations Research
MIT | Massachusetts Institute of Technology - Cambridge, MA, United States
Master's Degree in Operations Research
Columbia University - New York, NY, United States
Bachelor's Degree in Applied Mathematics and Economics (Double Major)
Yale University - New Haven, CT, United States
Certifications
NVIDIA DLI Certificate – Applications of AI for Anomaly Detection
NVIDIA
Skills
Libraries/APIs
Scikit-learn, Matplotlib, NumPy, Pandas, Ggplot2
Tools
Jupyter, Microsoft Access, BigQuery
Languages
Python, R, C++, Perl, SQL
Paradigms
Linear Programming, Anomaly Detection
Platforms
Linux, MacOS, Jupyter Notebook, Google Cloud Platform (GCP)
Storage
SQL Functions
Industry Expertise
Bioinformatics
Other
Machine Learning, Data Science, Artificial Intelligence (AI), Operations Research, Combinatorial Optimization, Graph Theory, Statistics, Network Optimization, Algorithms, Computer Science, Statistical Analysis, Statistical Modeling, Nonlinear Optimization, Economics, Credit Risk, Algorithmic Trading, Computational Biology, Chemistry, IT, Genomics, DNA Sequencing, Feature Engineering, Google BigQuery, Drug Development, OR-Tools, Fintech
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring