Much of my prior work has been proprietary however below is a listing of the generic technical and computational skills leveraged in these use cases.
Technical skills:
Languages: Python, R, SQL
IT knowledge and Software Experience: API integration, Flask, Docker, Tableau, MySQL Workbench,
Anaconda, Github, R studio, Jupyter Notebook, JSON development, Structured and Unstructured databases and datasets
Technical Specializations: Machine Learning, Deep Learning, Natural Language Processing, Graph Analysis, Neural Networks, Nearest neighbor/grouping cluster analysis, PCA, SVM,
integration Business Management Skills: Business Process Modeling, Agile Development, JIRA Utilization, Sprint Planning, Requirements Gathering, GitHub, Excel charting/pivot tables, Dashboarding, KPI development, MS Office.
Example solutions:
Sentiment prediction platform that surfaces news stories, events and their related sentiment, aggregating based on story similarity with layered data visualizations presented back to risk and trading teams in near real time.
> Google Cloud Platform (GCP) AutoML Natural Language; Natural Language API
> Topic modelling (LDA); similarity (cosine) modeling
> Python data scraping, pipeline development & GCP integration
> Tableau integration and front end development
Automated NLG and corporate credit risk platform & reporting solution, leveraging a 3rd party deterministic NLG solution and integrated with 3rd party data providers, run on company reported earnings:
> Narrative Science NLG configuration
> S&P API data pipeline & integration w/ JSON
> Python text data extraction from Moody’s industry PDFs
Predictive anti-money laundering service & ordinal ranking engine, integrated into a customer on-boarding and periodic review processes:
> Log odds and logistic regression ordinal modelling
> SQL data munging
Regulatory text classification & correlation custom client solution. I played more on the go to market strategy and client pitch side here while collaborating closely with our modeling & SME teams on the design and build.
> LSTM model for multi-classification model
> Document similarity through various similarity measures - cosine etc
> DAG data structures to organize regulation content ontology and relationships.
Movie and TV show NLP engine helping understand similar people, based on show endorsements, who are not connected as well as consider how commentary language can provide insight into show impact and viewer preferences.
> Topic modelling (LDA); similarity (cosine) modeling
> Python NRCLex and VAD analytics