Data Scientist - Natural Language Processing

Job Description :

A driven data scientist who will be able to partner with several internal and external stakeholders in order to deliver cutting edge client risk analyses. This position will also focus on R&D in enhancing current analysis and providing thought leadership for future studies. This position will address the needs of all industry groups manufacturing, financial services, energy, construction, healthcare, transportation, etc as they relate to all lines of commercial business across AIG.

Responsibilities include but not limited to:

- Help improve existing data analyses by always researching cutting edge statistical modeling and data visualization techniques.

- Help deliver customized analyses leveraging internal and external, structured and unstructured datasets. Utilize advanced modeling techniques to parse out adverse experience drivers. Manipulate large data sets, using cutting edge machine learning or statistical modeling techniques and synthesizing insights.

- Help develop an analytics platform to house and analyze big data

- Help with the process of curating, cleaning and integrating data to enable scalability of analyses. Leverage several internal databases as well as external data sources.

- Collaborate with other analytical teams to leverage tools and techniques developed across AIG.

- Always strive to automate analyses where possible.

- Help with advancing Data Visualization capabilities

- Help train new staff in a growing team. Help manage additional COE resources as needed.

Desired Candidate Profile

- The ideal candidate needs to be familiar with the following techniques and tools, with an expert-level experience in some:

- Expertise in one or more modeling/machine learning platforms as such as R, SAS, and Python

- Expertise in Natural Language Processing techniques

- Classification methods (e.g., Neural Net, Logistic Regression, Decision Trees, KNN, SVM, Random Forest)

- Regression methods (e.g., Linear, Nonlinear, Boosted Regression Trees ) Clustering methods (e.g., K-means, Fuzzy C-means, Hierarchical Clustering, Mixture Modelling)

- Time-series Modelling/Forecasting (e.g., AR, ARMA, GARCH, Exponential Smoothing)

- Statistical Analysis (e.g., Hypothesis Testing, Experiment Design, Hierarchical Modeling, Bayesian Inference)