← All speakers

Bio, Work & Ideas

Gabriela de Queiroz

Conference affiliation: Microsoft · 2024

Gabriela de Queiroz founded f02 labs, a developer-relations consultancy for AI companies, and R-Ladies, the international organization advancing gender diversity in statistical computing. Previously a chief data scientist at IBM and director of AI at Microsoft for Startups, she specializes in helping developers and founders gain the technical resources, practical guidance, and professional networks needed to build useful products.

Building communities, then AI programs

Raised in Brazil, de Queiroz studied statistics and epidemiology before continuing her statistics education in California. In 2012, she founded the first R-Ladies chapter in San Francisco. Four years later, she helped establish R-Ladies Global with support from the R Consortium, expanding a local meetup into an international network organized around technical education, mentorship, and professional opportunity.

She subsequently founded AI Inclusive, which supports gender minorities in artificial intelligence through workshops, educational programs, and scholarship partnerships. These organizations reflect her conviction that inclusive technical communities influence who acquires technical skills, builds professional confidence, and helps shape emerging technologies.

At IBM, de Queiroz held leadership positions in open-source technology, data, and AI before becoming a chief data scientist. She later joined Microsoft, leading educational advocacy across AI, machine learning, and data science before serving as director of AI at Microsoft for Startups.

At AI Engineer World’s Fair 2024, she helped founders navigate startup-support programs, open-source AI templates, and shared development environments. Her emphasis was concrete: GitHub Codespaces reduces inconsistent local setup, adaptable starter applications accelerate experimentation, and direct access to experienced advisers helps founders address architecture, hiring, and commercialization. Technical demonstrations involving retrieval systems and evaluation were led by other presenters; de Queiroz concentrated on founder support, customization, and making hands-on workshops genuinely usable.

After leaving Microsoft, she founded f02 labs, advising AI startups on positioning, community strategy, technical education, and developer adoption. Her engagements have included fractional developer-relations leadership at Galileo, an AI evaluation company.

  • Developer trust is earned through usefulness. Credible demonstrations, clear technical explanations, direct feedback, and established communities help AI products find audiences more effectively than promotional claims alone.
  • Access requires infrastructure and human support. Cloud resources, reproducible environments, mentors, scholarships, and peer networks remove different barriers; combining them gives newcomers and founders a stronger chance of making progress.
  • Practical AI education should produce working software. Hands-on workshops and customizable application templates turn unfamiliar tools into prototypes participants can test, adapt, and continue developing independently.

Read the topics behind these talks

2 conference talks

Key ideas

Scroll to read ↓

Build and deploy a Python chat app, add retrieval over database rows or documents, and follow the engineering decisions that turn a working demo into a testable application.

  • Getting an application running before building everything yourself
    1:25 ↗
  • Establish the account and model connection
    6:13 ↗
  • Open the template and understand its async backend
    17:39 ↗
  • Configure the connection and make the chat app your own
    23:21 ↗
  • Deploy with a separate environment configuration
    29:22 ↗
  • Recover from region and naming constraints
    37:48 ↗
  • Ground answers in retrieved database rows
    41:18 ↗
  • Turn documents into searchable, citable chunks
    46:04 ↗
  • Understand what changes when infrastructure changes
    52:52 ↗
  • Make quality observable and testable
    56:33 ↗
  • Inspect the failures behind retrieval scores
    1:03:29 ↗
  • Update embeddings deliberately, and retrieve code by structure
    1:09:38 ↗
  • Manage prompts and budget evaluation runs
    1:15:55 ↗
  • Separate model compatibility from reliable output
    1:20:53 ↗
  • Use computation when the question spans the whole dataset
    1:25:42 ↗

Key ideas

Scroll to read ↓

Build and deploy a Python chat application, then extend the same workflow to product retrieval and document answers with citations, query rewriting and evaluated search.

  • How do you get an AI application running quickly?
    1:13 ↗
  • Support beyond credits—and a shorter path to experimentation
    6:34 ↗
  • Separate deployment funding from model access
    11:46 ↗
  • Start with a shared development environment
    25:55 ↗
  • Run the chat app and ask for a haiku
    29:06 ↗
  • Use async I/O and stream the response
    33:01 ↗
  • Deploy through Azure Developer CLI and Bicep
    36:23 ↗
  • Retrieve product rows before generating an answer
    42:11 ↗
  • Turn conversational requests into search constraints
    45:49 ↗
  • Make documents searchable and answers inspectable
    48:05 ↗
  • Customize the app, then inspect what ingestion actually extracts
    51:18 ↗
  • Evaluate the retrieval combination, not just the vector index
    54:51 ↗
  • Confirm a working result and stop unused environments
    57:24 ↗
  • Keep the document index current
    1:01:12 ↗

References