top of page
  • GitHub Maira Gatti de Bayser
  • CNPq
  • LinkedIn - Black Circle

Maira Gatti de Bayser, PhD


Dr. Maira Gatti de Bayser former Research Staff Member, has co-founded the IBM Research Brazil, the Brazilian Lab of IBM T.J. Watson Jr. Research Center. She successfully technically led and coordinated top R&D projects in the Deep Learning/Machine area for more than 15 years. She pioneered on the establishment of the Conversational Systems area in Brazil before ChatGPT. That resulted in national and international publications, patents, and system demonstrations, for instance, at IJCAI’2019, and the premier league organization “HUMAINE (HUman Multi-Agent Immersive NEgotiation Competition) 2020” at IJCAI’2020, which happened in 2021 due to Covid. All the R&D Cloud-based Machine Learning solutions that she has led had international and national publications in top AI Conferences, like AAAI and IJCAI, patents, and business results. Maira is a member of ACM, AAAI, and ACS. Maira holds Ph.D. (2009) in Science - Informatics, and M.Sc. (2006) in Informatics from PUC-Rio, Pontifical Catholic University of Rio de Janeiro, Brazil. Her Ph.D. entitled Engineering of Self-Organizing Emergent Multi-Agent Systems: A Design Method and Architecture contains a novel method that was implemented using Monte-Carlo Simulation, a Planner and Reinforcement Learning. While her M.Sc. work entitled Dependability of Law-Governed Open Multi-Agent Systems had a novel method to replicate law-mediated agents. Maira has 70 published scientific papers and 30 US Patents in AI.

Professional CV: 

CarEer path
- 2022
Bachelor's Degree in Informatics
Institute of Mathematics and Statistics, UERJ
Software Engineer
@Internet Ventures
MSc in Informatics
PUC-Rio, in collaboration with Un. Pierre et Marie Curié
PhD in Informatics
PUC-Rio, University of Waterloo, King's College London
AI Research Software Engineer and Professor
PUC-Rio, Software Engineering Lab (LES), Computer Graphics Lab (Tecgraph)
Research Staff Member
IBM Research Brazil
Data Structures and Algorithms Professor
Visitant Professor, EMAp, FGV-Rio
Research Staff Member
IBM Research Brazil
Post-Graduated Program of Informatics
Collaborator Professor, Federal University of State of RJ
Patent US20220028366A1
Embodied Negotiation Agent and Platform
Patent US10070050B2
Device, system and method for cognitive image capture
Patent US9485319B1
Simulation to find user behavior impact in social media network
Patent US20170154111A1
Managing item life-cycle at home with internet of things
Patent US20130086462
Method and System for Retrieving Legal Data for User Interface Form Generation by Merging Syntactic and Semantic Constraints
April 2020
July 2019
Feb 2018
HUMAINE: Human Multi-Agent Immersive Negotiation Competition
2020 CHI Conference on Human Factors in Computing Systems, Tokyo, Japan.
Embodied Conversational AI Agents in a Multi-modal Multi-agent Competitive Dialogue
In the Proc. of the 28th International Joint Conference on Artificial, IJCAI 2019, Macau, China.
Specifying and Implementing Multi-Party Conversation Rules with Finite-State-Automata
32nd AAAI Conference on Artificial Intelligence (AAAI 2018), New Orleans, USA.
Ravel: a MAS Orchestration Platform for Human-Chatbots Conversations
Engineering Multi-Agent Systems (AAMAS 2018), Stockolm, Sweden. Integrated to IBM Watson Development Cloud.
Tutorial: Programming Conversational Agents with Akka Actors
The 18th European Agent Systems Summer School 2016

In this tutorial I present how conversational agents can be implemented as Akka Actors using a mixed-initiative dialog strategy (Gatti de Bayser et. al, 2017), i.e., by combining both the system- and user-initiative in a question and answer scenario. Akka is a toolkit and runtime that implements the Actor Model on the JVM and was inspired by Erlang in many ways. Akka can be used to build highly concurrent, distributed, and resilient message-driven applications. A large ecosystem of more than 250 Github projects has grown on Akka, in programming languages like Scala, Java, Groovy, and JRuby. The tutorial covers topics from agent pro-activeness, natural language processing, fault-tolerance, and finally an exemplar integration with a mobile app to demonstrate the Cognitive Investment Advisor Application, including the system and code.

Gatti de Bayser et. al, 2017: A Hybrid Architecture for Multi-Party Conversational Systems.
Jul 2015
Dec 2014
Deep Convolutional Neural Networks for Sentiment Analysis of Short Texts.
Int. Conf. on Computational Linguistics, COLING, Dublin, Ireland.
Aug 2014
Dec 2013
bottom of page