> Markdown version of [/videos/78-leverage-cloud-computing-benefits-with-serverless-multi-cloud-ml?t=1815](https://www.wearedevelopers.com/videos/78-leverage-cloud-computing-benefits-with-serverless-multi-cloud-ml?t=1815). 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).

---

# Leverage Cloud Computing Benefits with Serverless Multi-Cloud ML 

Raw data once caused an AI to classify a human as a juggling ball. Avoid this by chaining AWS, Azure, and GCP into a serverless machine learning pipeline.

- **Speakers:** Linda Mohamed
- **Event:** WeAreDevelopers LIVE
- **Published:** November 30, 2020
- **Duration:** 39:58
- **URL:** https://www.wearedevelopers.com/videos/78-leverage-cloud-computing-benefits-with-serverless-multi-cloud-ml

## Summary

Exploring the practical application of serverless multi-cloud machine learning, this session demonstrates how developers can build AI solutions without requiring deep mathematical or data science expertise. Using a creative prototype based on detecting and categorizing juggling patterns, the presentation unpacks the true democratization of AI. Instead of getting bogged down in complex decision trees, prototype builders can rapidly assemble managed services from major public cloud providers to bridge the gap between problem and practical solution. Because machine learning is fundamentally about data, the core challenge often lies in data collection and cleaning rather than algorithmic training. Initial manual attempts to label objects quickly prove unscalable, leading to the construction of an automated multi-cloud pipeline. A web scraper feeds internet imagery into an AWS S3 bucket, triggering AWS Lambda to pass data to Azure Machine Learning. Azure Custom Vision acts as a pre-processing filter to clean noisy user-generated content, ultimately formatting a refined dataset for Google Cloud AutoML to handle the primary object detection training. Deploying the final solution highlights the flexibility of containerization and the power of connecting disparate technology dots. By exporting the trained model as a Docker container, the architecture achieves true multi-cloud portability, allowing the model to run locally or across any public cloud without vendor lock-in. Key takeaways from this project emphasize that raw user-generated data will easily confuse models without robust pre-processing (which once led the AI to classify a human as a juggling ball), that serverless operations massively reduce infrastructure overhead, and that minimizing the time and specialized experience needed to build functional prototypes represents the ultimate value of cloud democratization.

**Keywords:** serverless multi-cloud architecture, machine learning pipelines, artificial intelligence democratization, automated feature engineering, unstructured data pre-processing, object detection models, azure custom vision, google cloud automl, aws lambda integrations, docker container deployment, cloud computing benefits, ai prototype development, computer vision services, public cloud ecosystems

## Chapters

1. **Understanding foundational data science and artificial intelligence concepts** (04:27) — Differentiating computer science, data science, and machine learning lays the groundwork for practical artificial intelligence implementation.
1. **Explaining the standard machine learning development cycle** (10:00) — Fetching, preparing, training, and evaluating data forms the core iterative loop of machine learning model generation.
1. **Evaluating existing pre-trained models and mathematical calculations** (12:43) — Existing open-source machine learning solutions and theoretical math often fail to address practical real-world variables.
1. **Manual data pre-processing using custom vision platforms** (15:54) — Training a custom vision model manually reveals the limitations and time constraints of non-automated data labeling.
1. **Exploring data collection challenges from user-generated platforms** (18:09) — Scraping visual data from social platforms highlights the need for robust data acquisition workflows.
1. **Automating feature engineering to handle messy data** (19:19) — Uncleaned social media inputs cause model recall metrics to drop, proving the necessity of comprehensive data cleaning.
1. **Connecting multi-cloud services for automated data preparation** (22:36) — Utilizing an event-driven flow across public cloud platforms scales machine learning data cleaning processes efficiently.
1. **Orchestrating serverless workflows with cloud messaging services** (25:37) — Cloud messaging queues and serverless functions format data automatically for cross-platform model training.
1. **Comparing cloud provider model training and evaluation accuracy** (27:56) — Contrasting model evaluation across different cloud vendors demonstrates minor performance differences in image recognition.
1. **Deploying machine learning models locally and remotely with containers** (30:15) — Packaging exported machine learning models into Docker configurations allows for seamless multi-cloud deployment.
1. **Defining tech democratization across serverless cloud environments** (33:43) — Leveraging multiple managed cloud infrastructure platforms empowers software engineers to bypass deep domain expertise requirements.
1. **Reviewing architectural prototype lessons and future deployment pipelines** (36:29) — Understanding that manual implementation prototypes lack maintainability emphasizes the requirement for automated continuous delivery pipelines.

## Related Moments

- [Exploring hybrid models, monoliths, and serverless computing](https://www.wearedevelopers.com/videos/261-why-you-shouldn-t-build-a-microservice-architecture) (from "Why you shouldn’t build a microservice architecture ")
- [Exploring the tiered architecture of modern machine learning stacks](https://www.wearedevelopers.com/videos/154-machine-learning-for-software-developers-and-knitters) (from "Machine Learning for Software Developers (and Knitters)")
- [Evaluating cost effectiveness and summarizing serverless machine learning strategies](https://www.wearedevelopers.com/videos/158-serverless-deployment-of-large-nlp-models) (from "Serverless deployment of (large) NLP models ")
- [Transitioning artificial intelligence infrastructure into scalable commodity cloud services](https://www.wearedevelopers.com/videos/1001-langchain4j-an-introduction-for-impatient-developers) (from "Langchain4J - An Introduction for Impatient Developers")
- [Comparing containers, cloud architectures, and serverless computing](https://www.wearedevelopers.com/videos/333-cd2cf-continuous-deployment-to-cloud-foundry) (from "CD2CF - Continuous Deployment to Cloud Foundry")
- [Practical learnings from deploying containerized AI solutions](https://www.wearedevelopers.com/videos/1127-using-containers-to-deploy-ai-models-across-our-microscopy-platform) (from "Using Containers to deploy AI Models across our microscopy platform")

## Related Articles

- [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)
- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [MLOps – What’s the deal behind it?](https://www.wearedevelopers.com/magazine/125-mlops-what-s-the-deal-behind-it)
- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)

## Related Jobs

- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/588393-machine-learning-engineer) at **Twilio**
- [Data Scientist](https://www.wearedevelopers.com/jobs/ext/1351648-data-scientist) at **Almedia**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/1355348-machine-learning-engineer) at **TWILIO**
- [Staff, Machine Learning Engineer (L4)](https://www.wearedevelopers.com/jobs/ext/1202639-staff-machine-learning-engineer-l4) at **Twilio**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**