Machine learning MLNG
(modified)
Developing systems that learn from data and experience, with the capability to improve performance over time.
SFIA 9 is in development
- SFIA 9 beta due in early July 2024
- SFIA 9 planned for publication October 2024
Moving to SFIA 9
- Guidance notes and level descriptions have been updated to maintain contemporary view of this skill including machine learning operations (ML Ops)
- SFIA 8 skill descriptions will remain available for you to use
- Previous SFIA assessments are not impacted by this change
Guidance notes
(modified)
Activities may include, but are not limited to:
- selecting and applying appropriate machine learning techniques, algorithms, and tools to solve business problems
- preparing and pre-processing data for machine learning tasks, including data cleaning, transformation, and feature engineering
- designing, training, optimising, and periodically retraining machine learning models using techniques such as supervised, unsupervised, or reinforcement learning
- evaluating trained models for performance, robustness, and bias, and selecting and using metrics to assess outcomes
- diagnosing and resolving issues before and after deployment
- anticipating the organisational implications of machine learning models, including ethics, bias, privacy, and data protection
- establishing traceability for the outcomes produced by machine learning systems
- implementing continuous learning mechanisms to ensure models adapt to new data and changing environments, including developing models that can adapt in real-time to new data inputs and evolving conditions.
Understanding the responsibility levels of this skill
Where lower levels are not defined...
- Specific tasks and responsibilities are not defined because the skill requires a higher level of autonomy, influence, and complexity in decision-making than is typically expected at these levels. You can use the essence statements to understand the generic responsibilities associated with these levels.
Where higher levels are not defined...
- Responsibilities and accountabilities are not defined because these higher levels involve strategic leadership and broader organisational influence that goes beyond the scope of this specific skill. See the essence statements.
Developing skills and demonstrating responsibilities related to this skill
The defined levels show the incremental progression in skills and reponsibilities.
Where lower levels are not defined...
You can develop your knowledge and support others who do have responsibility in this area by:
- Learning key concepts and principles related to this skill and its impact on your role
- Performing related skills (see the related SFIA skills)
- Supporting others with tasks (generic examples are provided by the essence statements for each level)
Where higher levels are not defined...
- You can progress by developing related skills which are better suited to higher levels of organisational leadership.
Levels
| Defined at these levels: | 2 | 3 | 4 | 5 | 6 |
Click to learn why SFIA skills are not defined at all 7 levels.
Show/hide extra descriptions and levels.
Level 1
Machine learning: Level 2
(modified)
Assists in data preparation, model training, and evaluation tasks under routine supervision.
Uses standard machine learning frameworks and tools to develop basic models for well-defined problems.
Documents results and contributes to the maintenance of machine learning solutions.
Machine learning: Level 3
(modified)
Applies established machine learning techniques and algorithms to solve business problems.
Selects and prepares appropriate data for model training and evaluation.
Trains, optimises, and validates machine learning models using standard tools and frameworks.
Deploys models into production and monitors their performance. Communicates results and limitations to stakeholders.
Machine learning: Level 4
(modified)
Assesses machine learning suitability and designs and develops machine learning solutions for a range of business problems.
Selects and applies appropriate techniques and algorithms based on data characteristics and project requirements. Provides guidance to others.
Engineers features and optimises model performance. Implements algorithms, and contributes to development, evaluation, monitoring, and deployment. Applies industry-specific rules and guidelines, anticipating risks and implications.
Collaborates with cross-functional teams to integrate machine learning models into production systems. Conducts in-depth performance analysis and troubleshoots issues.
Machine learning: Level 5
(modified)
Leads the development and implementation of machine learning solutions for complex, high-impact business problems.
Architects end-to-end machine learning pipelines and systems. Evaluates and selects appropriate tools, frameworks, and infrastructure for machine learning projects.
Establishes good practices and standards for machine learning development and operations. Provides expert advice and guidance on machine learning techniques and applications.
Collaborates with stakeholders to align machine learning initiatives with organisational goals.
Machine learning: Level 6
(modified)
Sets the strategic direction and roadmap for machine learning adoption and innovation within the organisation. Establishes governance frameworks and best practices for responsible and ethical development and use of machine learning.
Leads the development of organisational capabilities, policies, standards, and guidelines in machine learning.
Collaborates with senior stakeholders to identify high-impact opportunities for machine learning and drives their implementation.
Actively follows research and industry trends and integrates them into organisational practices.