Data Science

Specializations

Data Science Management Consulting and Advisory Services

Organizations that attempt to do data science at any scale beyond small experiments/focused projects often have multiple missteps, and sometimes total failures. Axiologic Solutions offers a series of consulting and advisory services to assist organizations to 1) optimize/improve how data science is currently being used, and 2) how to scale data science to the enterprise level.

The focus of the Data Science Management Consulting and Advisory Service includes:

  • Assessment (a la CMMI) of the maturity and consistency of how data science is being performed – the lifecycle process itself, the quality of the models, and analysis being produced
  • Assessment of the skills of the data scientists
  • How to establish a data science professional development program (or talent development program)
  • How to incorporate emerging trends in data science that may affect the organization
  • How to reduce the costs and increase the value/ROI of data science
  • How to approach a particular data science project (strategy, project planning, project startup, QA)

Establishing a Data Science capability.

Many government organizations are considering establishing a new data science capability or maturing an existing data science capability. As part of the Axiologic Solutions data science portfolio, we offer management consulting services on how to establish or improve an organizational level data science capability for a government agency, including creating and delivering a full-spectrum training and education curriculum.

Data Engineering.

As data becomes increasingly critical to a variety of business domains, including data science, business intelligence, cyber security, data mining, data migration, data analysis, and business reporting (including auditing), the need for well-thought-out and sound data solutions has emerged; this is commonly referred to as “data engineering.”

Though the term “data engineering” is relatively new, this function also has deep roots in systems engineering, and software engineering, particularly in ETL data processing, and data warehousing. Simply put, data engineering is focused on helping an organization and its many functions locate, access, move, process, and analyze data.

Axiologic Solutions personnel have the required expertise and experience to cover a wide range of data engineering usage scenarios in support of data science and other initiatives.

Advanced Machine Learning Techniques.

Axiologic Solutions provides specialized consulting services on how to perform advanced ML at enterprise scale, including:

  • Use of AutoML tools and techniques
  • Feature Engineering, including:
    • Representational feature learning: Feature Distributed Representations / Embeddings (GloVe, x2Vec)
    • Feature Dimensionality Reduction (SVD, PCA, t-SEN, Correlation Coefficient, LDA, Information Gain, Alpha-Test)
  • Incremental Learning and On-Line Learning, such as:
    • Learning Under Concept Drift
      Learning Under Data Drift
  • Transfer Learning
  • Privacy Preserving Learning
  • Security Classification Aware/Sensitive Learning
  • Veracity Sensitive Learning
  • Secure Learning / Adversarial Learning
  • Fair (Unbiased) Learning
  • Explainable Machine Learning
  • Federated Learning

Management of Data Science Projects.

Data science projects (DSP) are sufficiently different from other types of projects – such as software development, systems engineering, acquisition support, or management consulting – that they require different program/project management approaches.

DSPs are typically characterized by:

  • Incomplete or non-existent detailed requirements (in the traditional sense), which are replaced by high level business objectives or thesis
  • High failure rate to meet the business objectives; some research shows a DSP failure rate as high as 70-80%
  • Difficult to measure and control quality
  • Low reproducibility across different teams over time
  • Difficult to transfer and/or operationalize
  • Low maintainability over time
  • Difficult to determine completion criteria, often leading to “spiraling down the drain” work that consumes resources with no benefit .

Axiologic Solutions has synthesized the successful project management practices from dozens of leading data science organizations (e.g., Google, Facebook, Amazon, JP Morgan Chase, Visa) spanning many sizes, scopes, and industries to create the “gold standard” of management for DSPs, encompassing the complete data science life cycle.