Video by FINOS via YouTube
What happens when a business-focused professional discovers that AI can translate vision into working software?
🗽 Catch Host Peter Smulovics Live in New York! Ready to explore emerging technologies, spec-driven development, and the future of open-source financial engineering?
Join Peter Smulovics and global tech leaders at OSFF New York on November 4–5, 2026.
🎟️ Register Now: https://hubs.ly/Q04n_bZL0🔥 25% OFF DISCOUNT CODE: FINOSREDOSFF25
In this episode, we follow the journey of a business-aligned leader who never saw herself as a developer. Armed with curiosity, domain expertise, and increasingly sophisticated prompting techniques, she crosses the Rubicon from consumer of technology to creator of technology.
We explore how modern AI tools are redefining who gets to build software, contribute to projects, experiment with emerging technologies, modernize financial systems, and solve real-world business problems. The conversation examines what changes when technical barriers fall, and domain expertise becomes a superpower.
This isn’t a story about replacing developers. It’s a story about expanding the circle of creators. Host Peter Smulovics (Distinguished Engineer at Morgan Stanley and FINOS Zenith Lead) sits down with Fintech Product Manager Sarah Hilton-Burroughs for a special edition of the Open Source in Finance Podcast. Together, they explore how domain experts are leveraging AI agents, Spec-Driven Development (SDD), and custom component prompts to build full-stack RegTech platforms—without abandoning architectural guardrails or regulatory auditability.
🕒 Timestamps:
0:00 Welcome & Episode Introduction
0:32 Meet Sarah Hilton-Burroughs: 10 Years in Fintech & RegTech Product Management
1:53 The Spark: From Learning Python Curriculum to the Lightbulb Moment
3:36 Domain Expertise as Scripting: Why Writing Requirements Is the Ultimate Prompt
5:42 Moving Past Simple Prompts: Navigating Token "Credit Suck" & Multi-Tenant Stacks
7:27 Embracing Spec-Driven Development (SDD) & Customizing Component Agents
9:21 Misconceptions of Non-Technical Builders: Addressing AI Fabrication vs. Hallucination
12:24 The Game Boy Approach: Button-Masher Curiosity vs. Reading Manuals
13:48 Domain Expertise as a Competitive Moat in the Age of Commodity Code
16:09 How Financial Institutions Will Change: High-Value Use Cases vs. License Waste
17:57 Skills for the Future: Integrity, Data Science, and Regulatory Guardrails
22:12 AI-Native Builder Roles & The Reality of Commercial AI Affordability
25:09 User vs. Orchestrator: Navigating the LLM Spectrum (Sonnet, Opus, Cohere)
28:30 Preserving the Human Voice: Why Authenticity Beats AI-Generated Communications
33:51 What Being a "Developer Wizard" Means: Unimaginable Creative Freedom
35:45 Audience Q&A: Regulatory Auditability, Deterministic Lineage, and NIST/GDPR SDLC
📊 The Problem: The Context-Window "Credit Suck" & AI Fabrication
When non-technical domain leaders attempt to build complex software using basic chat prompts, they quickly encounter two major barriers: context decay and AI fabrication. Re-teaching an entire multi-tenant codebase to a general AI model on every iteration rapidly consumes token credits without maintaining state. Furthermore, unguided models fabricate plausible-sounding code and logic rather than adhering strictly to empirical laws or financial regulations, creating massive compliance risks if deployed blindly.
🏗️ The Solution: Component-Driven SDD & Specialist Agent Fleets
Sarah Hilton-Burroughs demonstrates how business leaders can structure their domain knowledge into a scalable software pipeline: Spec-Driven Prompt Libraries: Replacing endless conversational chats with structured markdown specification libraries and component-level rule sets. Specialized Agent Delegation: Customizing dedicated AI agents for specific architectural layers (e.g., a schema agent that exclusively manages database mutations). Domain Knowledge Moats: Recognizing that code writing has been democratized, shifting the builder’s focus to requirements precision, data science fundamentals, and cross-jurisdictional compliance.
⚙️ Why This Matters for Financial EngineeringEmbedded Regulatory Auditing:
Integrating NIST AI RMF, ISO, and GDPR frameworks directly into the SDLC to generate immutable data provenance, lineage, and mathematical state logs that pass strict regulatory scrutiny. Expanding the Creator Base: Empowering product managers and domain experts to rapidly prototype and validate solutions alongside senior system architects, drastically accelerating time-to-market for financial tools.
🌐 More about FINOS: https://www.finos.org/
📧 Join our newsletter: https://www.finos.org/sign-up
🎙️ Listen to our Open Source in Finance Podcast: https://www.youtube.com/@FINOS/podcasts
LinkedIn: https://www.linkedin.com/company/finosfoundation
#OSFFNewYork #ZenithSIG #PromptEngineering #SpecDrivenDevelopment #RegTech #AIGovernance #FinTech #OpenSource