Managing Artificial Intelligence Projects: Framework for Success. Part 8
People
Core team
The success of an AI project critically depends on having the right people involved. Each project must have individuals in these roles:
· Product Owner: Own the final product and ensure its success. Define requirements and goals, decide on scope, liaise with stakeholders, and represent the business voice.
· Project Manager: Onboard the team, monitor project progress, manage the budget, ensure key milestones are met, facilitate updates, manage risks, and lead iteration planning.
· Data Engineer: Transform raw data into usable formats for data scientists and stakeholders. Tasks include collecting, integrating, cleansing, and structuring data.
· Data Scientist: Solve business challenges using scientific methods and algorithms. Conduct exploratory data analysis, feature engineering, and AI algorithm development.
· Machine Learning / AI Engineer: Bridge the gap between testing AI models and production. Automate workflows, ensure proper data pipeline connections, optimize models for production, and focus on CI/CD.
Extended team
Depending on organization’s budget and resources, it is recommended having an extended team of stakeholders:
Project Sponsor: Initiates project, advocates, and provides budget and resources.
Cloud Architect: Translates technical requirements into cloud architecture; designs components for storage, compute, and tools.
Cloud Administrator: Manages platform; ensures proper function, conducts security and performance tests.
Subject Matter Expert (SME): Offers domain expertise in a specific business or technical area.
Business Analyst: Analyses data for patterns and insights; reports findings for informed decisions; bridges technical and business success metrics.
AI Expert: Develops machine learning models; specializes in AI fields like machine vision and natural language processing.
Regulatory/Quality Expert: Ensures product compliance with regulations; conducts audits and reviews deliverables during development.
Legal/Privacy Expert: Approves data usage and product release; drafts privacy policies and terms of use, facilitates smooth production transitions.
Certain stakeholders in AI projects often have overlapping responsibilities, leading to individuals filling multiple roles simultaneously. For instance, we might see a product owner also acting as a project manager, a Machine Learning Operations (MLOps) engineer doubling up as a data scientist, or a data engineer taking on a data scientist’s duties. Similarly, it’s not uncommon for a subject matter expert to serve as a data scientist, a cloud architect to function as a cloud administrator, or a product owner to fulfil business analyst tasks.
However, if one individual assumes more than two roles within a complex AI project, it raises a red flag. In such situations, the project sponsor must step in to ensure that additional resources are allocated, preventing burnout, and ensuring the project remains on track. For example, if a data engineer is asked to also handle data science and cloud administration responsibilities, the project sponsor should recognize this strain and bring in additional team members to distribute the workload.
REFERENCES
Amershi, S. et al. (2019) ‘Software Engineering for Machine Learning: A Case Study’, in 2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP). 2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP), Montreal, QC, Canada: IEEE, pp. 291–300. Available at: https://doi.org/10.1109/ICSE-SEIP.2019.00042.
Bernardi, L., Mavridis, T. and Estevez, P. (2019) ‘150 Successful Machine Learning Models: 6 Lessons Learned at Booking.com’, in Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. KDD ’19: The 25th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Anchorage AK USA: ACM, pp. 1743–1751. Available at: https://doi.org/10.1145/3292500.3330744.
Chinese University of Hong Kong et al. (2021) ‘Strategic Directions for AI: The Role of CIOs and Boards of Directors’, MIS Quarterly, 45(3), pp. 1603–1644. Available at: https://doi.org/10.25300/MISQ/2021/16523.
DeNisco Rayome, A. (2019) ‘Why 85% of AI projects fail’. Available at: https://www.techrepublic.com/article/why-85-of-ai-projects-fail/ (Accessed: 18 May 2024).
EU AI Act (2024). Available at: https://artificialintelligenceact.eu/the-act/.
Lou, B. and Wu, L. (2020) ‘Artificial Intelligence and Drug Innovation: A Large Scale Examination of the Pharmaceutical Industry’, SSRN Electronic Journal [Preprint]. Available at: https://doi.org/10.2139/ssrn.3524985.
Mead, C. and Ismail, M. (eds) (1989) Analog VLSI Implementation of Neural Systems. Boston, MA: Springer US (The Kluwer International Series in Engineering and Computer Science). Available at: https://doi.org/10.1007/978-1-4613-1639-8.
Universiteit Amsterdam et al. (2021) ‘When the Machine Meets the Expert: An Ethnography of Developing AI for Hiring’, MIS Quarterly, 45(3), pp. 1557–1580. Available at: https://doi.org/10.25300/MISQ/2021/16559.
University of Cologne et al. (2021) ‘Will Humans-in-the-Loop Become Borgs? Merits and Pitfalls of Working with AI’, MIS Quarterly, 45(3), pp. 1527–1556. Available at: https://doi.org/10.25300/MISQ/2021/16553.
Wessel, L. et al. (2019) ‘Configuration in smart service systems: A practice‐based inquiry’, Information Systems Journal, 29(6), pp. 1256–1292. Available at: https://doi.org/10.1111/isj.12268.
Wong, N. et al. (2012) ‘Artificial immune systems for the detection of credit card fraud: an architecture, prototype and preliminary results’, Information Systems Journal, 22(1), pp. 53–76. Available at: https://doi.org/10.1111/j.1365-2575.2011.00369.x.
By Yesbol Gabdullin, Senior Data Scientist, MSc in Artificial Intelligence, MRes in Engineering, BEng in Electronics and Control Engineering, Microsoft Certified
The article was written for the Forecasting the Future of AI Applications in the Built Environment session of the Project Controls Expo 2024.