Data Scientist

Position Description: 

Researches and develops Machine Learning (ML) models that identify suspicious transactions and customers using Artificial Intelligence (AI) and ML-based systems. Programs and develops computer systems with Python and SQL. Researches, develops, and delivers models to detect suspicious activity within typologies that include, but are not limited to, cryptocurrency trading, cryptocurrency receipt/delivery, market manipulation (equities, options, and cryptocurrencies), securities fraud, insider trading, elder financial exploitation, and money-laundering and terrorist financing domains 

Primary Responsibilities: 

• Performs exploratory analysis, data cleaning, preparation and annotation, ML pipeline design and development, model evaluation, and validation.  

• Develops models using supervised and unsupervised ML algorithms. 

• Develops models using algorithms, such as decision trees, isolation forests, autoencoders/neural networks, linear/logistic regression, and clustering.  

• Develops models that operate on both structured and unstructured data (Natural Language Processing). 

• Analyzes and preprocesses features for model training. 

• Collaborates with team to assess project scope, define data requirements, prioritize tasks, and share research findings and updates. 

• Researches new techniques and technologies to improve team knowledge and enhance solutions. 

• Creates presentations to provide team updates on project progress, research, and new findings. 

• Participates in code reviews to enable learning, collaboration, and mentoring of other team members. 

Education and Experience: 

Master’s degree (or foreign education equivalent) in Computer Science, Engineering, Information Technology, Information Systems, Mathematics, Physics, or a closely related field and no experience. 

Skills and Knowledge: 

Candidate must also possess: 

• Demonstrated Expertise (“DE”) performing with complex SQL queries to extract features from SQL databases; using Python language for typical DS workflow steps -- data preprocessing, regression, decisions trees/random forest, neural network, feature selection/reduction, clustering, and parameter tuning. 

• DE developing data pipelines on Amazon Web Services (AWS) using S3 storage services and EC2 Cloud computing services; performing supervised and unsupervised modeling on tabular data in Python, using DS libraries (Pandas, NumPy, SciPy, and Scikit-Learn); training models on imbalanced datasets using python libraries (IMBlearn); creating data visualizations to analyze and evaluate model results, using Python libraries (Matplotlib and Seaborn). 

• DE developing classification models on text data, using Spacy, NLTK, Tensorflow, Pytorch, and BERT frameworks. 

• DE communicating and collaborating across teams to break down complex business problems, translate into ML projects, and deliver data products and insights for productization, using collaboration tools (JIRA). 

#PE1M2