Curriculum Vitae

Hector focuses on AI problems that generate structured outputs based on data and/or business and contextual descriptions. He investigates AI algorithms that build upon existing solutions for simpler AI challenges. In applied projects, he leads the development of robust AI systems attuned to business contexts, ensuring they tackle the right problems while adapting to user contexts and preferences.

Basics

Name Héctor Palacios
Label AI Research Scientist
Email hectorpal@gmail.com
Phone 438.496.LS66
Summary Work in AI problems with structured outputs given data and/or description of business and context. Research in AI algorithms that use existing algorithms and components for simpler AI problems. Lead the development of robust AI aware of business contexts, engaging the right problem to solve.

Work

  • 2021.03 - 2023.08
    Staff Research Scientist
    ServiceNow Research
    Fundamental research in AI reasoning and NeuroSymbolic methods. Support applied research within ServiceNow.
  • 2018.09 - 2021.02
    Applied Research Scientist
    Element AI
    Reasoning on top of a ML model. Achieved double-digit absolute improvement in hard cases of a vision-related task, without losing good performing cases of the pure ML-based model.
  • 2016.11 - 2018.09
    Senior Research Scientist
    Nuance Communications
    Intent similarity by Siamese Network using dense data. Question answering on domain-specific content and no training data.

Education

  • 2003.10 - 2009.12

    Barcelona, Spain
    PhD
    Universitat Pompeu Fabra
    Computer Science

  • Caracas, Venezuela
    BS & MS
    Universidad Simón Bolívar
    Computer Science

Volunteer

  • 2008 - 2010
    Coordinator
    Online Activism
    Stimulated, coordinated and enabled online activism for data crowdsourcing during natural disasters and elections.

Awards

Skills

Programming
Python
Java
C/C++
Functional languages
Linux
Computer clusters
Docker
Version control
Research Coordination and Leadership
Defining problems and scope
Designing evaluations
Harnessing team skills and interests
Aligning strategically with research and business objectives

Languages

English
Professional
Spanish
Native
French
Elementary

Interests

Fundamental Research Problems
NeuroSymbolic approaches
Task-centric dialogue
Symbolic Planning
Preference fulfillment and optimization
Symbolic reasoning
Properties of Methods
Robustness
Interpretability
Systematic generalization