> Markdown version of [/videos/1127-using-containers-to-deploy-ai-models-across-our-microscopy-platform?t=602](https://www.wearedevelopers.com/videos/1127-using-containers-to-deploy-ai-models-across-our-microscopy-platform?t=602). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Using Containers to deploy AI Models across our microscopy platform

Tired of dependency hell when shipping AI models to client environments? Discover how containerizing workloads with Docker and WSL2 unlocks native GPU access and decouples your stack.

- **Speakers:** [Sebastian Rhode](https://www.wearedevelopers.com/@sebastian-rhode)
- **Event:** World Congress 2024
- **Published:** August 20, 2024
- **Duration:** 23:56
- **URL:** https://www.wearedevelopers.com/videos/1127-using-containers-to-deploy-ai-models-across-our-microscopy-platform

## Summary

Microscopy generates massive imagery datasets that require robust AI models to execute complex tasks like instance segmentation and cell classification. Historically, deploying these trained machine learning models to diverse client environments introduced significant friction, as cloud platforms and localized software needed identically synchronized dependencies to ensure reproducible results. To solve this, engineering teams modernized their pipelines by shifting the deployment artifact from a bare model file to a unified, containerized application. By wrapping the model alongside its pre-processing, post-processing, and data I/O within a Docker container accessed via a REST interface, they cleanly decoupled their AI stack from client application code. A critical enabler for this architecture is WSL2, which grants Linux containers native GPU access on Windows machines, allowing fast inferencing without sacrificing desktop performance. This separation of concerns significantly reduces technical debt and empowers AI developers to independently distribute updated algorithms without requiring disruptive client software patches. The migration to containers demonstrated clear business value by solving tangible issues around scalability and frictionless distribution, proving that technological adoption wins stakeholder approval when tied directly to real-world outcomes. For organizations pursuing similar modernizations, prototyping early with straightforward computer vision workloads safely derisks the technical transition, while relying on open container standards reliably eliminates vendor lock-in.

**Keywords:** ai model deployment challenges, microscopy computer vision, containerized machine learning dependencies, instance segmentation workflows, wsl2 gpu acceleration, docker desktop integration, reproducible ml pipelines, ai stack synchronization, decoupling client code, reducing technical debt, azure ml model training, open container standards, legacy risk mitigation, cross-platform analytics, business value metrics

## Chapters

1. **Introduction to Zeiss and microscopy technology** (00:02) — An overview of the Zeiss business units focused on industrial quality and research microscopy.
1. **Exploring common AI workflows in microscopy tasks** (01:21) — How computer vision handles classification and instance segmentation of biological samples.
1. **Managing large data sets in microscopy workflows** (02:40) — The transition from gigabyte to terabyte datasets when capturing time-lapse imaging of growing cells.
1. **Customer requirements for AI model delivery** (05:50) — The need for robust AI performance across cloud and local platforms without requiring IT expertise.
1. **Challenges in traditional AI deployment methods** (08:53) — Why decoupling the model from client software caused persistent and difficult dependency mismatches.
1. **Shifting to containerized AI deployment environments** (10:02) — Bundling models and all necessary dependencies into Linux containers ensures consistent execution across environments.
1. **The containerized model training and deployment pipeline** (12:40) — Generating a containerized artifact after cloud training drastically simplifies distribution to image analysis clients.
1. **Executing containerized AI in production workflows** (17:06) — Downloading and running the self-contained AI module locally integrates smoothly into the user perspective.
1. **Business value of independent AI module updates** (18:25) — Delivering improved AI features quickly via containers speeds up progress without risking client software stability.
1. **Practical learnings from deploying containerized AI solutions** (20:24) — Starting with prototypes and open container standards minimized deployment risk in a traditional engineering company.

## Related Moments

- [Optimizing and deploying containerized AI inference workloads](https://www.wearedevelopers.com/videos/920-wwc24-ankit-patel-unlocking-the-future-breakthrough-application-performance-and-capabilities-with-nvidia) (from "WWC24 - Ankit Patel - Unlocking the Future Breakthrough Application Performance and Capabilities with NVIDIA")
- [Standardizing AI model and tool execution using containers](https://www.wearedevelopers.com/videos/1384-compose-the-future-building-agentic-applications-made-simple-with-docker) (from "Compose the Future: Building Agentic Applications, Made Simple with Docker")
- [Adopting AI tools for developer container workflows](https://www.wearedevelopers.com/videos/100183-from-build-to-breach-hacking-kubernetes-through-the-supply-chain) (from "From Build to Breach: Hacking Kubernetes Through the Supply Chain")
- [Optimizing AI deployments with bootable containers](https://www.wearedevelopers.com/videos/1593-bootable-ai-containers-with-podman-desktop) (from "Bootable AI Containers with Podman Desktop")
- [Deploying machine learning models locally and remotely with containers](https://www.wearedevelopers.com/videos/78-leverage-cloud-computing-benefits-with-serverless-multi-cloud-ml) (from "Leverage Cloud Computing Benefits with Serverless Multi-Cloud ML ")
- [Bridging application containerization and modern AI agent development](https://www.wearedevelopers.com/videos/1384-compose-the-future-building-agentic-applications-made-simple-with-docker) (from "Compose the Future: Building Agentic Applications, Made Simple with Docker")

## Related Articles

- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [Building AI Solutions with Rust and Docker](https://www.wearedevelopers.com/magazine/494-building-ai-solutions-with-rust-and-docker)
- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)
- [MLops – Deploying, Maintaining And Evolving Machine Learning Models in Production](https://www.wearedevelopers.com/magazine/115-mlops-deploying-maintaining-and-evolving-machine-learning-models-in-production)

## Related Jobs

- [Head of AI Applications](https://www.wearedevelopers.com/jobs/ext/1231536-head-of-ai-applications) at **ZEISS Group**
- [Head of AI Applications](https://www.wearedevelopers.com/jobs/ext/1456210-head-of-ai-applications) at **ZEISS Group**
- [Security Architect - AI](https://www.wearedevelopers.com/jobs/ext/1581899-security-architect-ai) at **ZEISS Group**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/1597388-machine-learning-engineer) at **ZEISS Group**
- [Product Owner - Artificial Intelligence](https://www.wearedevelopers.com/jobs/ext/396346-product-owner-artificial-intelligence) at **ZEISS Group**
- [Senior Cloud DevOps Engineer](https://www.wearedevelopers.com/jobs/ext/1271985-senior-cloud-devops-engineer) at **ZEISS Group**