# S12 E14: Catalina Turlea, Lovelaice

Page: https://stenobird.com/podcast/code-story/s12-e14-catalina-turlea-lovelaice
Text version: https://stenobird.com/podcast/code-story/s12-e14-catalina-turlea-lovelaice.md
Podcast: [Code Story: Insights from Startup Tech Leaders](https://stenobird.com/podcast/code-story)
Published: 2026-04-14T10:00:28+00:00
Episode link: https://codestory.co/podcast/e14-catalina-turlea-lovelaice/
Audio file: https://pdst.fm/e/pscrb.fm/rss/p/audio4.redcircle.com/episodes/c2ac4221-1192-4e27-874c-b0193d09d9af/stream.mp3
Processing state: processed
JSON: https://stenobird.com/v1/public/podcasts/code-story/episodes/s12-e14-catalina-turlea-lovelaice
Duration seconds: 1409

## Resource

Catalina Turlea shares how her experience running a tech consultancy revealed a gap in how companies implement AI features. She explains the transition from identifying 'prompt-based' failures to building Lovelaice, a platform for validating AI product value.

## Highlights
- Main idea: Avoid the 'AI feature' trap where products rely on unoptimized prompts that fail to deliver user value
- Practical takeaway: Use evaluation frameworks including deterministic tests and 'LLM as a judge' to validate AI performance
- Failure mode: Building features based on founder assumptions rather than validated user pain points and real-world use cases
- Strategic insight: Prioritize product smoothness and core value over adding unnecessary technical complexity or 'cool' features
- Foundational advice: Validate that customers are willing to pay for a solution before committing significant engineering resources

## Topics

AI Product Development, LLM Evaluation, Startup Strategy, Serverless Architecture, Product Validation, Software Engineering Management, Founder Journey

## Chapters
- 1:00 — The Serverless Approach: Catalina discusses her philosophy of using existing infrastructure and AWS serverless to avoid reinventing the wheel.
- 5:30 — From Consultancy to Founder: Reflecting on 14 years of product building and how consulting for various startups sparked the idea for Lovelaice.
- 7:40 — Validating AI Features: How to use test datasets and multiple LLMs to find the best configuration without heavy engineering overhead.
- 10:00 — Prioritizing Product Smoothness: The importance of focusing on user experience and making features seamless rather than just adding new capabilities.
- 14:30 — The Evaluation Framework: Implementing error analysis and LLM-based judging as part of a continuous feedback loop with pilot customers.
- 19:10 — Scaling and Infrastructure: Lessons learned from scaling services from hundreds to hundreds of thousands of users using serverless architecture.
- 21:40 — Building Efficient Teams: Catalina's approach to leading a small, high-growth, and predominantly female engineering team.

## Actions

- request_transcript: `POST https://stenobird.com/v1/public/podcasts/code-story/episodes/s12-e14-catalina-turlea-lovelaice/transcription-requests` — Idempotently request low-priority transcript generation for this episode.
- read_markdown: `GET https://stenobird.com/podcast/code-story/s12-e14-catalina-turlea-lovelaice.md` — Read the agent-friendly Markdown representation of this episode resource.

A page view does not enqueue transcription. Agents should invoke `request_transcript` explicitly when they need this episode processed.

## Transcript

Full transcripts are not published on public pages unless there is a clear rights basis.