Now that you know the AIAP origin story and what we look for in an apprentice, this article will share a bit more about the AIAP programme.
Introduction
The AIAP has 2 months of deep skilling phase, followed by a 7-month project phase with the majority of apprentices working on 100E projects and some on internal projects such as developing AI products like Synergos, our federated learning framework, or MLOps tooling and LLM products in the AI Products team.
Deep Skilling Phase (2 Months)
The 2-months are intensive and often a shock to many apprentices as most apprentices would have self-taught themselves AI/ML algorithms, and very few would learn about docker containers, Git, good coding practice and software engineering in general before joining AIAP.
The curriculum covers machine learning fundamentals, advanced ML/deep learning, and MLOps focused on deploying an AI model. This 2-month there are no classrooms or lectures, but instead is structured as self-directed learning assignments and discussions with mentors. The curriculum is dynamic, and the AIAP AI Mentors will introduce new techniques and research papers as required.
Part 1 - Machine Learning Fundamentals
- Data acquisition and cleaning
- Exploratory data analysis
- Feature engineering
- Clustering models
- Dimensionality reduction techniques
Part 2 - Advanced Machine Learning and Deep Learning
- Fundamentals of neural networks
- Computer vision methods (CNNs, image preprocessing, etc)
- Natural language processing methods (word embeddings, LLMs, etc)
Part 3 - Deployment
- Containerization with Docker
- Testing and documentation
- Model deployment (web apps, APIs, etc)Apprentices get exposure to theory and hands-on practice across the ML workflow – from data to training to deployment. This builds a solid foundation before tackling real-world AI projects.
Project Phase (7 Months)
Most apprentices will work on real-world industry projects via the 100 Experiments (100E) and help companies solve business problems using AI, build minimum viable products and deploy AI models into production and experience working with stakeholders.
As 100E project sponsor companies need to co-invest at least $75,000 cash and $126,000 in-kind, the industry projects in the 100E program are real-world business problems the company really wants to solve and not a toy project the company is doing as “national service by providing internship projects for university interns”.
The AIAP project team consists of 1 full-time AI engineer as the mentor, 4-6 apprentices working full-time, a professional project manager overseeing 3-4 projects, and support from senior AI engineers and MLOps experts.
Unlike an internship model where trainees work alone at a sponsor’s office, the AIAP team works collaboratively at AI Singapore’s office. The apprenticeship model facilitates deep learning through close teamwork and guidance.
The full-time team works together through all project stages, from initial research to data processing, model development, MLOps implementation, deployment, and documentation. Daily interactions and agile sprints encourage intense learning.
The project manager coordinates across teams, work with the various stakeholders and handles administrative project tasks. The mentor provides hands-on guidance and shares technical expertise. Senior AI engineers and MLOps staff lend additional support.
This structured team model emphasizes guided learning and collaboration. Apprentices receive ongoing supervision rather than working independently. The immersive team experience accelerates skills development for real-world AI engineering.
After executing and delivering more than 70 projects, the below shows a typical 7-months project cycle.
Unlike a typical student internship with limited partners and part-time effort, the full-time AIAP team can fully immerse in the project.
The AI mentor and apprentices work standard office hours, Monday through Friday. This enables them to collaborate intensely and use production-grade frameworks to deliver the project.
With 4-6 dedicated apprentices guided by an experienced mentor, they can take on substantial real-world projects. The team has the time and support to follow industry best practices rather than cutting corners to meet tight deadlines.
The team develops solutions following the same standards used at tech companies and AI startups. This focus on quality, robustness, and engineering rigour accelerates the apprentices’ readiness for the industry. Some of the tools leveraged are shown below:
Through the deep skilling and project phases, AIAP develops well-rounded AI engineering skills.
Apprentices learn the theory, tools, and best practices used by industry professionals. The curriculum and team project experience ingrain software engineering rigour across the entire ML workflow.
Some of the key engineering best practices apprentices master include:
- Robust version control with Git and GitHub
- Comprehensive testing and validation
- Clear documentation for maintainability
- Modular, reusable code organization
- Proactive debugging techniques
- Automated deployment and infrastructure
- Consistent style compliance and code quality
- Effective collaboration and communication
The AIAP’s immersive structure provides end-to-end training in AI and proficiency in surrounding engineering disciplines. Apprentices graduate as skilled professionals grounded in proven software methodologies, ready to build, deploy and maintain real-world AI systems.
Conclusion
Through AIAP’s immersive structure, apprentices develop the multifaceted skills needed to deliver robust, real-world AI solutions. The deep skilling phase grounds apprentices in both AI techniques and software engineering best practices. Hands-on projects then reinforce a rigorous approach to building production-grade AI systems.
Rather than hypothetical academic examples, apprentices tackle business challenges that companies invest significant resources to solve. With guidance from experienced AI mentors, they work through the entire lifecycle of requirements gathering, data wrangling, model development and deployment. This end-to-end experience with real stakeholders accelerates learning beyond what any textbook or boot camp can provide. Apprentices learn by doing, absorbing lessons that stick with them into future careers.
Upon completing AIAP, graduates can seamlessly join technology teams with proven abilities to engineer reliable AI. With its intense immersion in real projects, guided mentorship, and focus on production readiness, AIAP prepares apprentices for impactful careers advancing AI for business and society.
No classroom, boot camp, or online training can deliver this.
Leave a Comment