Trilliant Health

Currently working on interesting and difficult business problems using machine learning, Data Version Control (DVC), PySpark, Azure, bootstrapped data, manually labeled data, weak supervision and lots of other useful tools with a great team of engineers, scientists, analysts, product managers, and business owners.

Data Scientist IV

Gridsearch Consulting

Data science, data engineering, machine learning, product development, data strategy, data science strategy, data up-skilling, python up-skilling and instruction, software engineering, app development, data science team planning, data science project analysis and feedback, establishment of data science best practices for your team, and data science interviewing.

Co-founder and Principal

Unum

• Lead various natural language processing (NLP) projects designed to more efficiently route and auto-resolve emails, delivering an improvement of more than 50% over existing solutions

Principal Data Scientist

• Lead the development of AI using natural language processing/understanding
• Developed an AI Governance Framework that mitigated unintentional bias in production models, independently audited DS solutions (both internal data scientists and external vendors), ensured robust code was shipped to production, and offered transparency to the business on how DS models were developed
• Mentored and managed junior data scientists

Senior NLP Data Scientist

Nashville Software School

Data Science Instructor - Assistant 

The General®

• Worked with multiple departments to ensure a successful implementation of multiple Data Science Claims projects that integrate with existing software
• Coordinated across the entire American Family enterprise to discuss global solutions to local data science problems
• Designed and developed Silhouette, a tool that automates the process of setting up a virtual environment in the cloud (AWS) that automatically configures version control (DVC/S3, Git/Bitbucket), standardizes directory structures for productionalizing models (Cookiecutter), and ensures reproducibility (Pipenv)
• Created, contributed to, and designed the company's first tech blog (www.autoregressed.com)

Lead Data Scientist

• Use natural language processing and machine learning to predict attorney involvement on claims which estimated yearly savings at $1.9M and optimized claims KPIs
• Work with multiple departments to ensure a successful deployment of the attorney prediction model to production and integration with an internal claims system
• Carefully craft test/control groups for projects to ensure validation of estimated monetary gains and other measures of success
• Utilize paragraph vectors (using gensim's doc2vec implementation) to perform concept detection to search for abstract concepts and ideas such as "Spanish speaking", "fracture", and "attorney representation" (among others) in claim notes 
• Implement online learning so as to update machine learning models in real time with new, unseen information
• Convert PDFs to text documents using OCR to be used by machine learning models 
• Create a RESTful web app to serve and label random internal claims images for easier manual labeling (using Falcon)
• Coordinate with key business touch-points to ensure the successful delivery of projects

Data Scientist II

• Perform various natural language processing tasks (R/tidytext, Python/gensim), including sentiment analysis, topic modeling, and Word2Vec/Doc2Vec
• Build a nation-wide price elasticity model using a Mixed Effects/Hierarchical Logistic Regression algorithm (with cross classified and nested random effects) to predict demand for a product at various pricing structures (R, Julia)
• Predict a claim's route by converting claim notes to Word2Vec word embeddings and feeding those word vectors into machine learning models for multi-class prediction (Python/Gensim/Scikit-learn)
• Build an interactive geo-spatial smoothing app (Python/Flask, PostGIS, Leaflet.js) for various key performance indicators using AWS S3 as the data store
• Collaboratively design and implement statistical techniques and machine learning models with team members and department heads
• Create pipeline to transfer data from SQL Server database on local server to PostgreSQL database on AWS EC2 instances
• Assist the data science team in transitioning from local servers to cloud computing on AWS
• Architect and implement RESTful APIs (Python/Flask, Celery, Redis, SQL) that expose various statistical models for consumption

Data Scientist I

Perception Health

•	Built machine learning models (Python) that used lab data to predict if a patient would be admitted or not and visualized results using an ROC curve and confusion matrix
•	Worked alongside Perception Health's Chief Data Scientist to analyze physicians’ referral patterns and consult hospitals on how to best reduce out-of-network referrals
•	Used GIS (PostGIS, Leaflet.js) techniques to visualize healthcare trends geographically
•	Lead production on Perception Health's flagship product since March 2015 to the present
•	Built customizable, interactive dashboards (Leaflet.js, D3.js, Crossfilter.js) into the primary platform
•	Optimized SQL queries and data preparation for faster graph rendering and page-load times
•	Utilized Agile methodologies (JIRA) to organize releases while ramping up features
•	Worked with the lead architect to optimize SQL scripts across the primary product's platform
•	Converted the platform to a RESTful API for more efficiency and to decouple the server-side code from the client side templates
•	Designed and developed an internal SQL application (Python, Flask) for non-programmers to easily upload data into a database, making repetitive communication with the database faster and more memory efficient

Software Engineer Project Lead

Software Engineer

Relode, LLC

• Designed and wrote programs that efficiently scraped useful data from various websites using Python and created web applications to be used internally by Relode sales representatives to access this data.
• Uploaded and maintained captured data in a MySQL database.
• Wrote Python scripts that reformated and correctly structured raw, unstructured data.

Software Developer (Contractor)

Digital Reasoning

• Parsed corrupted XML files in Python using ElementTree and Regular Expressions in order to pin point the corruption
• Designed and wrote programs that efficiently cleaned and structured the corrupt data in order to increase the accuracy rate of the firm’s predictive models

Checkd.in

• Employed web-parsing techniques to convert unstructured, raw data into a workable format to perform statistical analyses
• Utilized Facebook's Social Graph API to scrape useful data
• Designed and wrote programs that parsed and statistically analyzed Facebook data using Python

Software Developer (Intern)

UBS Wealth Management

• Conducted Valueline and Morningstar research on various firms to track position of securities over time.
• Mapped out clients’ financial future based on current investments and economic position.

Wealth Management Intern

Building data products @ Gridsearch | Principal DS @ Trilliant Health

In both my day job and consulting practice (Gridsearch Consulting), I partner with businesses to bring data-centric products and solutions to market. 

I also build machine-learning apps with and for my sons. Currently, I’m building a Natural Language Understanding app I call LADNER (Locating Assignments and Dates using Named Entity Recognition) that locates homework assignments and due dates on his teachers’ websites and emails, and then automatically adds them to my Trello calendar daily so we don’t miss homework assignments or quizzes. 

Before that, I built a machine-learning app called Achoo that predicted my son's inhaler usage using air quality, weather, and allergen data.

In my career and hobbies, I build machine learning and data-centric software that makes life easier.

Taplio

💡Timothy Dobbins💡's Linkedin Analytics

💡Timothy Dobbins💡

What is 💡Timothy talking about?

Who is engaging with 💡Timothy