Module 1: Introduction to MLOps and the ML Lifecycle

• Defining MLOps: Principles, goals, and key challenges in production ML
• The three stages of MLOps maturity: Manual, Automated Pipeline, and CI/CD
• The complete ML development lifecycle: Data Preparation, Model Training, Deployment, and Monitoring
• Comparison between DevOps and MLOps: Unique requirements for data and models
• Practical session: Setting up a foundational MLOps environment and toolchain (e.g., Git, virtual environment, cloud sandbox)

Module 2: Version Control and Experiment Tracking

• Versioning ML code, configuration, and dependencies using Git
• Data versioning concepts and tools (e.g., DVC, Delta Lake)
• Tracking ML experiments: Parameters, metrics, and models using tools like MLflow or Weights & Biases
• Model registry and centralized management of approved production models
• Auditing and lineage: Tracing model predictions back to the training data and code

Module 3: ML Model Packaging and Containerization

• Model serialization formats (e.g., Pickle, ONNX, joblib) and portability
• Introduction to Docker: Creating isolated, reproducible environments for ML services
• Writing efficient Dockerfiles for model inference endpoints
• Strategies for shrinking image size and securing containerized models
• Practical session: Containerizing a trained Scikit-learn or PyTorch model into a lightweight Docker image

Module 4: CI/CD for Machine Learning

• Continuous Integration (CI): Automated testing of code, data validation, and model sanity
• Continuous Delivery (CD): Automated promotion of models through staging environments
• Triggering automated pipelines based on code commits or model performance improvements
• Implementing testing strategies for ML (unit, integration, and performance testing)
• Practical session: Building a CI/CD pipeline using GitHub Actions, GitLab CI, or Jenkins to automatically build and test a model service

Module 5: Model Deployment Strategies

• Deployment patterns for real-time inference: REST APIs and serverless functions
• Deployment patterns for batch inference: Scheduled jobs and ETL integration
• Advanced deployment techniques: Blue/Green, Canary releases, and A/B testing
• Rollback mechanisms: Designing systems to safely revert to previous model versions
• Practical session: Implementing a Canary deployment strategy using a service mesh or API gateway to test a new model version with a small traffic percentage

Module 6: Monitoring and Logging in Production

• Logging best practices for ML services: Structured logs and centralized log aggregation (e.g., ELK stack)
• Collecting real-time inference metrics: Latency, throughput, and error rates
• Business metrics monitoring: Tracking the impact of model predictions on key performance indicators (KPIs)
• Setting up effective alerting and notification systems for model degradation or infrastructure failure
• Practical session: Integrating Prometheus and Grafana to visualize inference metrics and set up alerts for high error rates

Module 7: Infrastructure and Orchestration (Kubernetes/Cloud)

• Introduction to Kubernetes (K8s) for model orchestration and scaling
• Deploying model services on a K8s cluster using YAML manifests and Helm charts
• Managing autoscaling (HPA) for fluctuating inference traffic
• Using cloud-managed ML platforms (e.g., Azure ML, GCP Vertex AI, AWS SageMaker) for simplified operations

• Practical session: Deploying the containerized model from Module 3 onto a Kubernetes cluster and testing horizontal pod autoscaling

Module 8: Model Governance and Regulatory Compliance

• Establishing an auditable Model Card or Model Fact Sheet for every deployed model
• Compliance requirements: Understanding regulations (e.g., GDPR, ethical AI guidelines) and their impact on ML systems
• Implementing access control and security measures for model APIs and data pipelines
• Ensuring model fairness and bias detection throughout the MLOps lifecycle
• Practical session: Creating a Model Card artifact detailing the model's performance, fairness metrics, and training lineage

Module 9: Data and Model Drift Detection

• Concepts of data drift, concept drift, and model decay in production
• Statistical methods for detecting data drift in input features
• Techniques for detecting concept drift in target variables and residuals
• Automated retraining strategies based on drift detection and performance degradation
• Practical session: Implementing a drift detection service using a tool like Evidently AI to monitor feature distributions and trigger an automated alert

Module 10: Advanced MLOps Patterns and Future Trends

• Serving models at the edge and on mobile devices (TinyML, TensorFlow Lite)
• Developing and deploying feature stores for consistent online and offline feature serving
• MLOps for specialized domains: LLMs and Generative AI deployment challenges
• Introduction to advanced workflow orchestrators (e.g., Kubeflow Pipelines, Airflow)
• Practical session: Designing a high-level architecture for a real-time feature store integration into an existing model serving pipeline

Requirements:

·       Participants should be reasonably proficient in English.

·       Applicants must live up to Armstrong Global Institute admission criteria.

Terms and Conditions

1. Discounts: Organizations sponsoring Four Participants will have the 5th attend Free

2. What is catered for by the Course Fees: Fees cater for all requirements for the training – Learning materials, Lunches, Teas, Snacks and Certification. All participants will additionally cater for their travel and accommodation expenses, visa application, insurance, and other personal expenses.

3. Certificate Awarded: Participants are awarded Certificates of Participation at the end of the training.

4. The program content shown here is for guidance purposes only. Our continuous course improvement process may lead to changes in topics and course structure.

5. Approval of Course: Our Programs are NITA Approved. Participating organizations can therefore claim reimbursement on fees paid in accordance with NITA Rules.

Booking for Training

Simply send an email to the Training Officer on training@armstrongglobalinstitute.com and we will send you a registration form. We advise you to book early to avoid missing a seat to this training.

Or call us on +254720272325 / +254725012095 / +254724452588

Payment Options

We provide 3 payment options, choose one for your convenience, and kindly make payments at least 5 days before the Training start date to reserve your seat:

1. Groups of 5 People and Above – Cheque Payments to: Armstrong Global Training & Development Center Limited should be paid in advance, 5 days to the training.

2. Invoice: We can send a bill directly to you or your company.

3. Deposit directly into Bank Account (Account details provided upon request)

Cancellation Policy

1. Payment for all courses includes a registration fee, which is non-refundable, and equals 15% of the total sum of the course fee.

2. Participants may cancel attendance 14 days or more prior to the training commencement date.

3. No refunds will be made 14 days or less before the training commencement date. However, participants who are unable to attend may opt to attend a similar training course at a later date or send a substitute participant provided the participation criteria have been met.

Tailor Made Courses

This training course can also be customized for your institution upon request for a minimum of 5 participants. You can have it conducted at our Training Centre or at a convenient location. For further inquiries, please contact us on Tel: +254720272325 / +254725012095 / +254724452588 or Email training@armstrongglobalinstitute.com

Accommodation and Airport Transfer

Accommodation and Airport Transfer is arranged upon request and at extra cost. For reservations contact the Training Officer on Email: training@armstrongglobalinstitute.com or on Tel: +254720272325 / +254725012095 / +254724452588