Title of the talk/workshop
Design-Driven AI Development: From Design to Production Code with AI
Abstract of the talk/workshop
AI coding assistants and coding agents are changing the way software is developed. However, simply asking AI to generate code can quickly lead to inconsistent architecture, unclear boundaries, technical debt, and code that does not necessarily reflect the original design intent.
Design-Driven AI Development takes a different approach: instead of treating AI as a tool that simply generates code, we treat software design and architecture as the foundation that guides AI throughout the development process.
In this session, we will explore how requirements, architecture, domain models, API contracts, user journeys, and other design artifacts can be used as structured context for AI-assisted development.
The session will demonstrate a practical development flow:
Requirements → Design → Architecture → AI-assisted Implementation → Testing → Review
We will look at:
- Why prompt-driven coding approaches often struggle with larger and more complex systems
- What Design-Driven AI Development means in practice
- How software architecture and design decisions can guide AI coding agents
- Using design artifacts as context for generating implementation
- Generating Python services, APIs, tests, and supporting components from design
- Keeping generated code aligned with architectural boundaries and design intent
- Using AI agents across different stages of the software development lifecycle
- How developers and architects can remain in control while AI generates more of the implementation
- Common pitfalls, limitations, and lessons learned when adopting AI-native development practices
The session will include a live practical walkthrough showing how a software requirement can be transformed into a design and then implemented with AI, including the generation and validation of Python code and tests.
The goal is not to teach a particular AI coding tool, but to demonstrate an engineering approach to AI-assisted development that can help teams move faster without giving up architecture, quality, maintainability, and engineering discipline.
Category of the talk/workshop
5. Data Science, Machine Learning, and AI
Secondary categories:
8. Developer tools and automation
9. Web development
Duration (including Q&A)
60 minutes
Proposed structure:
- 10 minutes — The problem with prompt-driven AI coding
- 15 minutes — Design-Driven AI Development: concepts and approach
- 25 minutes — Practical live demonstration
- 10 minutes — Q&A
Level of Audience
Intermediate
The session is aimed primarily at software developers, senior engineers, technical leads and architects. Developers with basic Python knowledge will be able to follow the practical demonstration.
Speaker Bio
Santosh MVS is a technology and engineering leader with 20+ years of experience in software engineering, backend systems, APIs, distributed architectures, enterprise platforms, and technology transformation.
He has a strong hands-on engineering foundation across backend development, APIs, microservices and distributed systems, combined with extensive experience in modernizing legacy platforms and transforming them into modern, scalable and API-driven architectures.
His experience also includes CRM platforms, enterprise application architecture and large-scale technology modernization. His current area of interest is the evolution of software engineering through AI-assisted development, AI-native engineering and agentic software development.
Santosh is particularly interested in how architects and engineers can use AI to accelerate software delivery while maintaining architectural integrity, engineering quality and long-term maintainability.
Company/College: Amdocs Development Center India LLP
Email: manapragada.santosh@gmail.com
Years of Experience: 20+ years
Prerequisites(if any)
No specific software setup is required for the session.
Attendees should ideally have:
- Basic software development experience
- Familiarity with APIs and backend development
- Basic understanding of software architecture
- Basic Python knowledge is helpful, but not mandatory
The session will focus on engineering practices, architecture and AI-assisted development, rather than teaching Python syntax or a specific AI coding tool.
Title of the talk/workshop
Design-Driven AI Development: From Design to Production Code with AI
Abstract of the talk/workshop
AI coding assistants and coding agents are changing the way software is developed. However, simply asking AI to generate code can quickly lead to inconsistent architecture, unclear boundaries, technical debt, and code that does not necessarily reflect the original design intent.
Design-Driven AI Development takes a different approach: instead of treating AI as a tool that simply generates code, we treat software design and architecture as the foundation that guides AI throughout the development process.
In this session, we will explore how requirements, architecture, domain models, API contracts, user journeys, and other design artifacts can be used as structured context for AI-assisted development.
The session will demonstrate a practical development flow:
Requirements → Design → Architecture → AI-assisted Implementation → Testing → Review
We will look at:
The session will include a live practical walkthrough showing how a software requirement can be transformed into a design and then implemented with AI, including the generation and validation of Python code and tests.
The goal is not to teach a particular AI coding tool, but to demonstrate an engineering approach to AI-assisted development that can help teams move faster without giving up architecture, quality, maintainability, and engineering discipline.
Category of the talk/workshop
5. Data Science, Machine Learning, and AI
Secondary categories:
8. Developer tools and automation
9. Web development
Duration (including Q&A)
60 minutes
Proposed structure:
Level of Audience
Intermediate
The session is aimed primarily at software developers, senior engineers, technical leads and architects. Developers with basic Python knowledge will be able to follow the practical demonstration.
Speaker Bio
Santosh MVS is a technology and engineering leader with 20+ years of experience in software engineering, backend systems, APIs, distributed architectures, enterprise platforms, and technology transformation.
He has a strong hands-on engineering foundation across backend development, APIs, microservices and distributed systems, combined with extensive experience in modernizing legacy platforms and transforming them into modern, scalable and API-driven architectures.
His experience also includes CRM platforms, enterprise application architecture and large-scale technology modernization. His current area of interest is the evolution of software engineering through AI-assisted development, AI-native engineering and agentic software development.
Santosh is particularly interested in how architects and engineers can use AI to accelerate software delivery while maintaining architectural integrity, engineering quality and long-term maintainability.
Company/College: Amdocs Development Center India LLP
Email: manapragada.santosh@gmail.com
Years of Experience: 20+ years
Prerequisites(if any)
No specific software setup is required for the session.
Attendees should ideally have:
The session will focus on engineering practices, architecture and AI-assisted development, rather than teaching Python syntax or a specific AI coding tool.