About

Daniel Griffin is a Senior Data Scientist at Cisco on the Digital Lifecycle Journeys data science team. He has an Undergraduate in Computer Engineering from the University of Cincinnati, and a Masters in Computer Science with a focus on Artificial Intelligence and Machine Learning from the University of Wisconsin Madison. While at the University of Wisconsin Madison, Daniel taught a course on Artificial Intelligence to over 400 students, and worked on research initiatives as a part of the David C. Page Medical Machine Learning Laboratory. He has over 7 years of experience in the data science field, and well over 50 data science projects to his name. He has experience performing advanced R&D development of machine learning algorithms, managing data science projects, and developing deployment pipelines and architectures for data science services in the cloud. His work spans across multiple research and industry institutions such as Wright Patterson AFB Research Laboratories in Dayton Ohio, Sandia National Labs in Albuquerque New Mexico, and now Cisco.

At his current workplace Cisco, Daniel has created and lead various data science initiatives for the Customer Experience organization. Some of the major projects include a ‘Social Media AI’ initiative for applying advanced text processing machine learning algorithms to Cisco’s social media channels in an automated fashion, which has already performed around 70 million predictions. Another initiative called ‘Customer Journey AI’ focuses on predictive and causal models to drive metrics around the customer journey Cisco cares about. He currently focuses on an initiative called the ‘Prescriptive Actions Engine’ which uses advanced agent based AI algorithms to prescribe actions Cisco can take to improve a customer’s experience.

General Data Science Analysis Methods

Data Science Methods Commonly Used (I won’t list all of the algorithms because that would be too much. But, suffice it to say that I am familiar with a wide variety of algorithms/methods within each topic):

• Supervised
  • Every major method you can likely think of for classification, regression, structured prediction, etc… (including the classical and advanced frequentist and Bayesian statistical methods, and beyond into gaussian process regression, VAE, etc…), and a wide variety of less well known methods, or custom developed methods.
• Unsupervised
  • Outlier detection
  • Unsupervised clustering
  • Recommendation Systems and Matrix Completion Methods
  • Semi-supervised clustering with EM
  • And many others…
• Semi-Supervised
  • Transfer Learning using both data augmentation methods, model augmentation methods, and model sharing methods, exploration in zero-shot and one-shot learning, weak learning and distant supervision
• Reinforcement Learning
  • Model based, Learning Based, MC, TD(lambda), On/Off policy based, Approximation Based, Bayesian Methods, Deep Learning Based, Non-standard reward function based, etc…
• Stats Based
  • Parametric/Non-Parametric statistical estimation/modeling, advanced graphical statistical modeling, causal inference, GLM/GAM, GPR, VAE, Bayesian Methods for Prediction and Parameter Inference, Hypothesis Testing & HDI Bayesian Methods, ANOVA, MLE/MAP and Risk based methods.
• Time Series
  • Dynamic Bayes Nets, Gaussian Processes, Temporal Point Processes, ARIMA Models, RNN/LSTM/GRU Recurrent Networks, Sequence Mining, Classification, etc…

Specialtiy Data Science Methods Knowledge Areas

• Advanced statistical methods including those in Bayesian statistics and Causal Inference
• Advanced reinforcement learning
• Advanced natural language processing methods (classical and machine learning driven)
• Machine learning based approaches to time series analysis (Dynamic Bayes Nets, Temporal Point Processes, Gaussian Processes)

Practical Data Science Knowledge Areas

• Data Science Project Lifecycle and Management Knowledge (I’ve successfully developed and deployed tons of models into production environments)
• Cloud Based Machine Learning Development and Deployment Best Practices
• Distributed algorithm implementations to scale machine learning algorithms

Previous Work Institutions

• Cisco
• Wright Patterson Air Force Base Research Laboratories (AF Research Headquarters)
• Sandia National DOE Research Laboratories (Major DOE Research Installation)
• Novel Device Laboratory Research (University of Cincinnati)
• David C. Page Medical Machine Learning Laboratory (University of Wisconsin Madison)
• Intuit (Mountain View, CA)