Model Online / On-Site
Duration
Modules 5 modules

Course Objective & Scope

Git versions source code effortlessly, but gigabytes of images, audio, Parquet files and model weights do not belong in a Git repository. When a production model misbehaves or an audit occurs, answering "Which exact dataset snapshot and label revision trained this model?" requires dedicated data and model versioning.

This course covers content-addressable storage, DVC (Data Version Control) over S3/MinIO, zero-copy branching and merging at object-storage scale with LakeFS, and table time-travel with Delta Lake and Apache Iceberg. Participants design an auditable lineage architecture that binds code, data, annotations and model weights under a single immutable Git tag.

Who Should Attend?

  • MLOps and data engineers managing large structured or unstructured (image, text, audio) training datasets
  • Data scientists who need reliable rollback, reproducibility and experiment isolation
  • Teams building AI infrastructure in regulated sectors such as finance, healthcare, defence and manufacturing
  • Prerequisite: Solid Git workflows, Python and basic familiarity with object storage (S3/MinIO/Blob)

Key Takeaways

  • Version large datasets and model weights on S3/MinIO with DVC while keeping lightweight pointers in Git
  • Create isolated zero-copy data branches and atomic merges/rollbacks over object storage using LakeFS
  • Bind Apache Iceberg or Delta Lake snapshot IDs directly to model training metadata
  • Synchronise evolving annotation exports from Label Studio or CVAT with underlying raw-data versions
  • Implement an end-to-end audit trail that traces any production model back to its exact code, data snapshot and evaluation metrics

Curriculum

The Versioning Problem in ML and Storage Fundamentals
  • Why code, raw data, annotations, environments and model weights evolve at different cadences
  • Content-addressable storage (MD5/SHA256), deduplication and object-storage (S3/MinIO) mechanics
Data and Artefact Versioning with DVC
  • .dvc pointer files, remote storage setup, dvc push/pull and shared team caches
  • Cryptographically locking data transforms and training stages with dvc.yaml and dvc.lock
Git-Like Operations on Object Storage with LakeFS
  • Zero-copy branching, committing, merging and instant rollbacks across millions of objects
  • Using pre-commit and pre-merge hooks to block corrupted or schema-violating data from reaching main
Tabular Time-Travel and Annotation Versioning
  • Reconstructing exact historical training splits via Apache Iceberg and Delta Lake snapshot IDs
  • Versioning evolving annotation exports (JSON/COCO) alongside immutable raw media in active-learning loops
Model Packaging, Semantic Versioning and Audit Lineage
  • Safe weight formats (Safetensors, ONNX), semantic versioning for ML and immutable release bundles
  • Querying end-to-end lineage from a live production endpoint back to Git SHA, data snapshot and training run

Other Courses in This Category

  • Enterprise Model Registry and Model Lifecycle Governance

    Central Model Registry architecture, packaging standards, @champion/@challenger promotions, automated validation gates, model cards and security scanning, for teams governing how models move into production.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Model Serving and High-Performance Inference Architectures

    Online, batch and streaming inference patterns, Triton Inference Server, BentoML, Ray Serve and vLLM, dynamic batching, gRPC/REST, caching and load testing, for teams serving models at low latency and high throughput.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • CI/CD and Continuous Training (CT) for Machine Learning

    Validating data, models and code in GitHub Actions/GitLab CI, automated PR evaluation reports, Continuous Training (CT) triggers with Airflow/Argo/Prefect and GitOps deployment, for MLOps teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Container and Kubernetes Deployment for AI and ML Workloads

    Optimised Docker images for CUDA/PyTorch, Kubernetes NVIDIA GPU Operator, MIG/time-slicing, serverless/autoscaling inference with KServe and distributed training with Kubeflow/Volcano, for platform and MLOps teams.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • GPU Acceleration and Inference Optimization

    GPU memory/compute architecture, profiling with Nsight Systems, TensorRT, FlashAttention, KV cache management, AWQ/GPTQ/FP8 quantization and speculative decoding, for engineers cutting deep learning and LLM inference costs.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View
  • Enterprise MLOps and Model Deployment on Amazon SageMaker

    SageMaker Studio, Processing and Training Jobs, SageMaker Pipelines, Model Registry, Real-Time/Serverless/Async Inference, Model Monitor and IAM/VPC security, for teams running end-to-end ML on AWS.

    MLOps & AI Platform Engineering Corporate
    Live Course
    5 modules
    View

Get in touch