The document discusses improving developer experience in cloud-native environments through platform engineering and AI, emphasizing the need for automated, scalable, and self-service digital platforms. It highlights the complexities of modern software development and identifies bottlenecks that hinder productivity, advocating for centralized management and streamlined workflows to enhance efficiency and innovation. Additionally, it introduces key metrics for measuring the effectiveness of internal developer platforms (IDPs) and the use of retrieval-augmented generation (RAG) to enhance AI responses.

1. Migliorare la Developer Experience
in un mondo Cloud Native
con Platform Engineering e AI RAG
Mich Murabito
Developer Advocate Mia-Platform
CNCF Ambassador & DevNetwork Board Member

2. Who am I?
I’m passionate about community and public speaking.
I founded the DevelopersLife community, where I actively
contribute.
Developer Advocate
Hello! I’m Mich Murabito
I also engage in projects and
initiatives in the field of
sustainability and open source,
with the goal of creating positive
impacts and sharing my passion for
technology.
Board Member
Feel free to add me on LinkedIn

4. The ”H”" in
Kubernetes stands
for Happiness
― quoted by anonymous feat Mich Murabito
4

5. Kubernetes
is just
the tip of
the
iceberg
The Cloud Native realm dazzles with its diversity and complexity: over 1261
Cards in the CNCF landscape, hundreds of tools, and an ever-growing ecosystem
showcase a vibrant multifaceted nature. 5

6. The “B” in Development stands for Bliss
– if only bottlenecks didn't exist.
― quoted by anonymous feat Mich Murabito
6

7. Bottlenecks,
potentially
everyone's
development
hurdle.
Bottlenecks are critical points that reduce flow and productivity, such as
lengthy onboarding, extended approval processes, or inadequate resources,
slowing down product release. 7

8. Why Software Development Complexity Has Increased
03
8
Todayʼs software must be scalable, secure,
and interoperable, meeting user expectations
across a global network.
⬢ Scalability and Distribution
⬢ Data Growth
⬢ Integrations and Interoperability
⬢ User Expectations and UX
⬢ Security and Privacy
⬢ Automation and CI/CD
⬢ AI and Machine Learning
⬢ Complex Architectures
⬢ Collaboration and Distributed Teams
⬢ Continuous Evolution

9. ⬢ Scalability and Distribution
⬢ Data Growth
⬢ Integrations and Interoperability
⬢ User Expectations and UX
⬢ Security and Privacy
⬢ Automation and CI/CD
⬢ AI and Machine Learning
⬢ Complex Architectures
⬢ Collaboration and Distributed Teams
⬢ Continuous Evolution
Why Software Development Complexity Has Increased
03
9
Todayʼs software must be scalable, secure,
and interoperable, meeting user expectations
across a global network.

10. Standard practices
without optimizations
a.k.a We could do better,
but for today we'll go the “old-fashioned wayˮ

11. Standard practices without optimizations
Developer
Need a new Environment
01 Environments
Resource configuration
02 Create
Ops
Configuration of environment
(variables, secrets, etc)
Network setup
Security setup
Many others…

12. Developer
Need a new Environment
01 Environments
Resource configuration
02 Create
Ops
Configuration of environment
(variables, secrets, etc)
Network setup
Security setup
Many others…
● Jira
● Trello
● Asana
● etc etc
Standard practices without optimizations

13. Platform Engineering
A brief and gentle introduction

14. Platform Engineering
“It is the approach aimed at
designing and building highly
automated and scalable internal
digital platforms that provide
developers with a self-service
environment to accelerate the
software lifecycle. The goal is to
simplify infrastructure management
and optimize operational efficiency,
enabling companies to adopt
cloud-native architectures and
technologies, while also improving
Developer Experience DevX) and
ensuring rapid innovation.ˮ

15. Platform Engineering
● Automated and scalable platforms
● Self-service
● Simplify infrastructure management
● Adopt cloud-native architectures
and technologies
● Improving Developer Experience
DevX) and ensuring rapid
innovation.

16. Platform Engineering: Enabling Efficiency and Innovation
01 Centralization
Manage all projects
from a single place
02 Automation
Industrialize and
automate the entire
DevOps cycle
03 Efficiency
Avoid bottlenecks
due to organization
04 Self-Service
Provide necessary
technology in a
self-service mode
05 Simplicity
Solve complexities
associated with the
cloud
06 Quality
Improve the quality
and reliability of the
produced software

17. Centralizing development operations
into a single portal represents a strategic
innovation for significantly improving the
developer experience.
Centralization
Manage all projects from a
single place
17

18. Automation brings efficiency and
innovation, transforming complex
processes into simple and precise
operations.
Automation
Industrialize and automate
the entire DevOps cycle
18

19. Providing essential technology in a self-service
mode allows developers to independently access
the tools they need, enhancing efficiency and
autonomy in their work.
Self-Service
Provide necessary
technology in a self-service
mode
19

20. Tackling and solving cloud-related complexities
with a simple and straightforward approach,
streamlining the management and use of cloud
technologies.
Simplicity
Solve complexities
associated with the cloud
20

21. Overcoming organizational slowdowns by
implementing efficient solutions that promote
a smooth and uninterrupted workflow.
Efficiency
Avoid bottlenecks due to
organization
21

22. Focusing efforts on continuous improvement of
software quality and reliability, ensuring
high-performing and secure products.
Quality
Improve the quality and
reliability of the produced
software
22

23. Cooperation between developers, DevOps professionals, and platform engineering teams is crucial. These three
roles work in close synergy, merging their specific skills to design, implement, and manage innovative solutions.
Collaboration Between Developers, DevOps, and
Platform Engineering Team

24. IDP Through the Eyes
of a Developer
IDPs revolutionize the developer
experience by centralizing access
and management of development
environments, code repositories,
and CI/CD pipelines. They provide a
unified interface for managing the
application lifecycle, from coding to
deployment, significantly reducing
operational complexity and enabling
developers to focus on creating
quality code, rather than system
configurations and infrastructure
management.

25. IDP Through the Lens
of DevOps
IDPs transform the DevOps realm by
offering centralized control over
deployment pipelines, infrastructure
automation, and monitoring. They
enable DevOps professionals to
effortlessly orchestrate complex
environments, ensuring continuity,
security, and scalability. With IDPs,
DevOps can manage the entire
application lifecycle more efficiently,
reducing release times and
enhancing system resilience.

26. Platform Engineering
team Bridge Between
Dev and DevOps
The Platform Engineering team
leverages IDPs as a bridge to unite and
empower the efforts of Developers
and DevOps professionals. Through
the IDP, the team provides a common
substrate that enables both groups to
maximize their respective tasks. This
synergy allows Developers to focus on
code creation and innovation, while
DevOps can streamline deployment
and infrastructure management,
ensuring a smooth, secure, and
efficient workflow.

27. Platform Engineering: a few practical examples
⬢ On-Demand Resource Provisioning: Developers can
create new environments in a few clicks without
waiting for the DevOps team.
⬢ Rapid Experimentation: A developer can test new
features in isolated environments and easily
integrate them if they work.
⬢ Automated Security and Compliance: Every
environment created automatically follows security
policies without delays.
⬢ Automatic Scalability: The platform adjusts
resources in real-time during traffic spikes, with no
manual intervention.

28. Application A
Development
Teams

29. Application A
Development
Teams
Application B Application C Application D

30. Application A
Development
Teams
Application B Application C Application D
Common
Aspects

31. Application A Application B Application C Application D
Platform
A Platform is a common place where people share value.
Software Life Cycle
Team Rules
Infrastructure
Toolchain
Services
Configuration

32. Interaction
Capabilities
Platform
Create
Curate
Resources
Consume Resources

33. Product Teams
Product A Product B Product C Product D
Infrastructure Providers, Complex Systems, External SaaS, Cloud Service

34. Product Teams
Product A Product B Product C Product D
Internal Developer Platform (IDP)
Platform
Interfaces
Infrastructure Providers, Complex Systems, External SaaS, Cloud Service
Platform
Interfaces

35. Product Teams
Product A Product B Product C Product D
Internal Developer Platform (IDP)
Platform
Interfaces
Platform
Orchestrator
Infrastructure Providers, Complex Systems, External SaaS, Cloud Service

36. Product Teams
Product A Product B Product C Product D
Internal Developer Platform (IDP)
Platform
Interfaces
Platform
Orchestrator
Infrastructure Platform
Execution Environment: K8s, Container Managed Service, Functions
Infrastructure Providers, Complex Systems, External SaaS, Cloud Service

37. Product Teams
Product A Product B Product C Product D
Internal Developer Platform (IDP)
Platform
Interfaces
Platform
Orchestrator
DevOps Platform
Git Pipelines Testing
Monitoring Tracing Secrets & Security
Infrastructure Platform
Execution Environment: K8s, Container Managed Service, Functions
Infrastructure Providers, Complex Systems, External SaaS, Cloud Service

38. Product Teams
Product A Product B Product C Product D
Internal Developer Platform (IDP)
Platform
Interfaces
Platform
Orchestrator
DevOps Platform
Git Pipelines Testing
Monitoring Tracing Secrets & Security
Infrastructure Platform
Execution Environment: K8s, Container Managed Service, Functions
Infrastructure Providers, Complex Systems, External SaaS, Cloud Service
CLOUD ENGINEERS
Deliver and operates the execution environment
PLATFORM
ENGINEERS
Deliver
and
operates
the
platform-as-a-product
SOFTWARE
ENGINEERS
Develop
and
operates
applications

39. Metrics
a.k.a “If we have data, letʼs look at data.
If all we have are opinions, letʼs go with mineˮ

40. Metrics are a powerful tool to assess the effectiveness and impact of new Metholodies and Technologies,
providing detailed insights into performance, organizational efficiency and team productivity.
Metrics are the invisible engine behind every
modern innovation

41. By analyzing key metrics, organizations can identify areas for improvement and continuously
optimize processes, aligning operations with strategic goals.
User
Goal: Measure the
platform's impact on
Developer Experience
and productivity.
● Environment
provisioning time
● Ease of use
● Developer feedback
scores
● Active users
● Retention with Net
Promoter Score NPS
Metrics by stakeholders
Organization
Goal: Assess how the IDP
supports organizational
efficiency and delivery
speed.
● Time to Market
● Resource Utilization
● Cost Savings
● Deployment Frequency
● Change Failure Rate
Platform
Goal: Monitor platform
stability and team
effectiveness.
● Mean Time to
Resolution MTTR
● Platform uptime
● Continuous feature
development
● Code Commit
Frequency
● Code Review Time

42. Developed by the DevOps Research and Assessment DORA) team, with the first metrics released
in 2014, the DORA framework helps quantify software delivery and operational performance.
DORA Metrics
● Deployment Frequency: Measures how often the team
deploys code to production.
● Lead Time for Changes: Tracks the time from commit to
production deployment.
● Change Failure Rate: Percentage of deployments that
lead to production issues.
● Mean Time to Recovery MTTR: Average time taken to
recover from production failures.
Goal: To balance the need for rapid development and release
cycles with system stability, these metrics became pivotal in
demonstrating DevOps' impact on productivity and quality.

43. Introduced in 2021 by researchers at GitHub and Microsoft, the SPACE framework
provides a holistic way to assess developer productivity
SPACE Metrics
● Satisfaction and Well-being: Measures developer happiness
and job satisfaction.
● Performance: Evaluates the volume of work completed,
considering quality and value.
● Activity: Tracks team activity, such as commit frequency and
pull request counts.
● Communication and Collaboration: Observes the
effectiveness of team interactions.
● Efficiency and Flow: Analyzes process efficiency and
workflow smoothness, aiming to minimize interruptions.
Goal: To address productivity in a multi-dimensional way,
recognizing that a high-velocity environment also requires
developer well-being and sustainable practices for long-term
success.

44. AI and RAG
A brief and gentle introduction
Skip →

45. Retrieval-Augmented
Generation
“aka RAG is a technique that
combines advanced language
models LLMs with the ability to
retrieve information from external
sources in real-time. This allows
for generating more accurate and
up-to-date responses by
integrating the model's output
with specific and relevant data.ˮ

46. Retrieval-Augmented
Generation
● Combines LLMs
● With external sources
● Real-time
● Generating more accurate
responses

47. RAG
RAG
L
L
M
G
e
n
A
I
D
L
M
L
A
I
G
P
T

48. RAG
LLM Input
External Data
Fine-Tuned LLM backbone
Hallucination
RAG

49. Generate Embeddings

50. Generate Embeddings
⬢ Embeddings convert words or phrases into
numerical vectors that represent their
meaning in a multidimensional space.
⬢ During training, the model learns to position
similar concepts close to each other in the
embedding space, based on context.
⬢ Embeddings are used to compare words or
phrases, enabling semantic search and
language understanding by AI models.

51. Generate Embeddings
Vector DB
LLM
Embeddings

52. Generate Embeddings
Vector DB
The CRUD service is
You can use CRUD to…
CRUD can be created…
LLM
Embeddings

53. Generate Embeddings
Vector DB
0,3,1,2,4,1 …]
1,4,1,2,2,1 …]
3,3,5,2,2,1 …]
LLM
Embeddings

54. Vector DB
What is a CRUD
Service?
0,3,1,2,4,1 …]
Generate Embeddings
LLM
Embeddings

55. Use Rag

56. Use Rag
⬢ RAG generates embeddings of a query to
represent its meaning in a vector space
⬢ These embeddings are used to search for
relevant documents or information in a
database, finding items with similar
vectors in the embedding space.
⬢ Once the most relevant information is
retrieved, RAG uses it to enrich the
context and improve the model's
response generation.
⬢ By combining information retrieval with the
generative model, RAG provides more
accurate and up-to-date answers
compared to using the language model
alone.

57. Vector DB
Search Related:
0,3,1,2,4,1 …]
Use Rag
LLM
Embeddings
What is CRUD
Service?

58. Vector DB
Search Related:
0,3,1,2,4,1 …]
Use Rag
LLM
Embeddings

59. Vector DB
The CRUD is…
You can use CRUD to…
CRUD can be created…
Use Rag
LLM
Embeddings

60. Vector DB
The CRUD Service
is a Plugin that…
Given these docs:
The CRUD is…
You can use CRUD to…
CRUD can be created…
Answer the following question:
What is a CRUD Service?
LLM
Embeddings
Use Rag

61. The DEMO
The DEMO  Generate Embeddings

62. The DEMO  Generate Embeddings
Vector DB
LLM Embeddings
AI RAG
Template
AI RAG
Template
Chat
Open AI
MongoDB

63. The DEMO  Chat BOT

64. The DEMO  Chat BOT
Without RAG enabled
With RAG enabled

65. Vector DB
AI RAG
Template
AI RAG
Template
Chat
MongoDB
The DEMO  Chat BOT
LLM Embeddings
Open AI

66. Demo
Skip →

67. Another step forward...
We've only seen "docs", but what happens if we add other information?
● Information about my platform
● Architecture and features of the platform.
● Information about my applications
● Information on data managed by the Platform
● Real-time logs and error

68. Another step forward...
We've only seen "docs", but what happens if we add other information?
AI
/
RAG
Platform

69. Another step forward...
We've only seen "docs", but what happens if we add other information?
AI
/
RAG
Platform

70. Another step forward...
We've only seen "docs", but what happens if we add other information?
AI
/
RAG
Platform

71. Another step forward...
We've only seen "docs", but what happens if we add other information?
AI
/
RAG
Platform
Do you detect
specific errors on...?

72. Another step forward...
We've only seen "docs", but what happens if we add other information?
AI
/
RAG
Platform
How can I aggregate
this data?

73. Another step forward...
We've only seen "docs", but what happens if we add other information?
AI
/
RAG
Platform
How can I manage
this situation in my
project?

74. Practical Tips
This is an example of how to edit
this new template
INTRODUCTION

75. Essential Guidelines for a Better Developer Experience
1. Invest in a effortless self-service experience for developers.
2. Align technology closely with business objectives.
3. Support gradual habit changes with patience and guidance.
4. Encourage collaboration to avoid common pitfalls.
5. Creating solutions that people genuinely enjoy using.

76. Presentation Template: SlidesMania
Images: Generated with Open-AI and
Midjourney
Fonts: Roboto Mono
76
Credits
Images: Generated with Open-AI and Midjourney
Fonts: Roboto Mono

77. 77
Resources
● Platform Engineering Trends:
tag-app-delivery.cncf.io/blog/proposal-
platform-engineering-/
● Blog Mia-Platform:
mia-platform.eu/blog/
● CNCF AI WhitePaper:
cncf.io/reports/cloud-native-artificial-intel
ligence-whitepaper/
● AI-RAG-Template OSS (by Mia-Platform):
github.com/mia-platform/ai-rag-template
github.com/mia-platform/ai-rag-template-chat
Feel free to add me on LinkedIn

78. -Feedback
-Q&A
-Photo
If you have any comments or suggestions about my presentation, please do not hesitate to fill the
form. Every piece of feedback is valuable!
78
https://bit.ly/feedback-mich

79. Thanks
→ LETʼS KEEP IN TOUCH
Michel Murabito – Developer Advocate
michel.murabito@mia-platform.eu