Victor Dibia

Research & Engineering Lead (Agent Optimization, Developer Experience), Microsoft Core AI · Mountain View, CA
AI agents, developer tools, and human-AI interaction
Download CV (PDF)

PDF updated Thu Aug 20 2026

Bio

I am an experienced technical leader and Principal Research Software Engineer at Microsoft Core AI, where I build AI agent infrastructure and developer tools (creator of Agent Optimizer in Foundry Agent Service; core developer for AutoGen, now 60K+ stars, and Microsoft Agent Framework). My research interests span AI agents, developer tools, and human-AI interaction. My research has been published at conferences such as ACL, EMNLP, AAAI, and CHI (1,750 citations, 9 patents) and has received multiple best paper awards. My work has also been featured in outlets such as the Wall Street Journal, VentureBeat, CIO . I am an IEEE Senior member.
I hold a PhD in Information Systems from City University of Hong Kong (recipient of the HKPFS scholar award by the Hong Kong Research Grants Council). My dissertation studied developer contribution behaviour in software crowdsourcing contests - factors influencing participation, the impact of incentives on participation behaviour and the problem solving process within crowdsourcing contests. Prior to City University, I studied at the Information Networking Institute at Carnegie Mellon University where I earned a Masters degree in Information Networking. I previously worked as a Principal Research Engineer at Cloudera Fast Forward Labs, Research Staff Member at IBM Research, Technical Lead for MIT Global Startup Labs, Researcher at the Innovation Management Lab, Athens Information Technology Athens Greece, and founder/lead developer for a small startup focused on West African markets.
In my free time, I enjoy music (some self taught guitar, piano), sports (participated in the Hong Kong Marathon thrice) and exploring art (some drawing, 3d modelling, african mask art).

Experience

Microsoft Core AI

May 2025 – Present

Research & Engineering Lead (Agent Optimization, Developer Experience) · Mountain View, CA

Microsoft Research - AI Frontiers Lab

Oct 2021 – May 2025

Principal Research Software Engineer · Mountain View, CA
  • Core developer for AutoGen (52K+ stars, 5.9M+ downloads). Creator of AutoGen Studio (EMNLP 2024). Co-authored Magentic-One, industry reference architecture for autonomous multi-agent systems.
  • Created LIDA (3.2K stars), an automated visualization framework using LLMs (ACL 2023). Approaches adopted by Excel, Fabric, PowerBI teams and Project Sophia.
  • Improved offline evaluation metrics for GitHub Copilot (14% higher correlation with customer satisfaction) used to select models for millions of users (ACL 2023).

Cloudera Fast Forward Labs

Jan 2019 – Sept 2021

Principal Research Engineer · New York / Palo Alto
  • Led research reports on Deep Learning for Question Answering, Anomaly Detection, and Image Analysis. Built NeuralQA, an extractive QA library using BERT.
  • Led development of Applied ML Prototypes (AMPs) for Cloudera ML — became the standard tool for customer onboarding.

IBM Research

Apr 2016 – Jan 2019

Research Staff Member · Yorktown Heights, NY
  • Created Data2Vis, the first and most cited neural network approach to automatic data visualization (IEEE CG&A Best Paper, IEEE VIS Best Paper Honorable Mention).
  • Co-created TJBot, an open-source DIY AI kit adopted by 8,000+ users for classroom teaching and corporate training.

Selected work

All projects
Product
Agent Optimizer
Automatic evaluation and optimization of hosted AI agents in Microsoft Foundry Agent Service.

2026

Announcement

Agent Optimizer

Automatic evaluation and optimization of hosted AI agents in Microsoft Foundry Agent Service.

Agent Optimizer is a service in Microsoft Foundry Agent Service that automates the agent improvement cycle through a closed-loop process: it evaluates baseline agent performance against defined criteria, generates optimized candidate configurations, tests and ranks candidates by score, and deploys winning configurations with a single command. It supports multiple optimization targets - instructions (rewriting system prompts), skills (generating reusable procedures), models (evaluating across LLM deployments), and tool descriptions (improving function-calling clarity). An 'eval init' command generates test datasets and scoring criteria from existing agent instructions, addressing the cold-start problem, and a portal interface supports browsing optimization runs and comparing candidates. Announced at Microsoft Build 2026 and available to hundreds of thousands of customers on Azure.

2026

Announcement
Open
Open source
Microsoft Agent Framework
An open-source SDK and runtime for building, orchestrating, and deploying AI agents - the unification of Semantic Kernel and AutoGen.

2025 · 13K+ stars

GitHubDocs

Microsoft Agent Framework

An open-source SDK and runtime for building, orchestrating, and deploying AI agents - the unification of Semantic Kernel and AutoGen.

Microsoft Agent Framework is an open-source SDK and runtime (Python and .NET) for building, orchestrating, and deploying AI agents and multi-agent workflows. It unifies Semantic Kernel and AutoGen into a single framework and is integrated into Azure Foundry Agent Service. I led the unification effort and designed the middleware/guardrails architecture that defines interception points for all function, tool, and agent calls across the SDK.

2025 · 13K+ stars

GitHubDocs
Open
Open source
AutoGen Studio
A low-code interface for rapidly building, testing, and sharing multi-agent solutions.

2024

GitHubDocs

AutoGen Studio

A low-code interface for rapidly building, testing, and sharing multi-agent solutions.

AutoGen Studio is a low-code interface built on AutoGen, enabling developers to rapidly build, test, deploy, and share multi-agent solutions. It provides a user-friendly interface to create and customize agents with little to no coding required. AutoGen Studio allows users to rapidly author agent workflows via a user interface, interactively test and debug agents, reuse artifacts, and deploy workflows. Key features include the ability to choose from pre-defined agents, compose them into teams (workflows), customize agents with foundation models, prompts, and skills, and deploy workflows as APIs. Future plans include a drag-and-drop interface for workflow authoring and a community gallery for sharing workflows, agents, and skills.

2024

GitHubDocs
Open
Open source
LIDA
Automatic Generation of Grammar-Agnostic Visualizations and Infographics.

2023 · 3.2K+ stars

Project Page

LIDA

Automatic Generation of Grammar-Agnostic Visualizations and Infographics.

Systems that support users in the automatic creation of visualizations must address several subtasks - understand the semantics of data, enumerate relevant visualization goals and generate visualization specifications. In this work, we pose visualization generation as a multi-stage generation problem and argue that well-orchestrated pipelines based on large language models (LLMs) and image generation models (IGMs) are suitable to addressing these tasks. We present LIDA, a novel tool for generating grammar-agnostic visualizations and infographics. LIDA comprises of 4 modules - A SUMMARIZER that converts data into a rich but compact natural language summary, a GOAL EXPLORER that enumerates visualization goals given the data, a VISGENERATOR that generates, refines, executes and filters visualization code and an INFOGRAPHER module that yields data-faithful stylized graphics using IGMs. LIDA provides a python api, and a hybrid user interface (direct manipulation and multilingual natural language) for interactive chart, infographics and data story generation.

2023 · 3.2K+ stars

Project Page
Open

Education

City University of Hong Kong
PhD in Information Systems (Quantitative User Behaviour, HCI)

2012 – 2016

Carnegie Mellon University
MSc Information Networking

2009 – 2011

Awards

  • Designing Multi-Agent Systems — Amazon Best Seller, #1 New Release in AI

    2025

  • Best Paper Award, IEEE Computer Graphics & Applications

    2020

  • Grand Prize, #BuiltWithTensorflow Challenge

    2019

  • Best Paper Honorable Mention, IEEE VIS

    2018

  • Best Technical Demo, AAAI

    2018

  • Heidelberg Laureate Forum — 1 of 200 young researchers invited

    2018

  • IBM Open Source Award

    2017

  • Google Developer Expert, Machine Learning

    former

  • Google Cloud Certified Professional — Data Engineer, Cloud Architect

    former

Patents

7 granted · 2 pending

Sort
  • Automated generation of data visualizations and infographics using large language models and diffusion models
    US 12,518,447
    granted

    2026

  • Data Health Evaluation Using Generative Language Models
    US 12,579,115
    granted

    2026

  • Using Large Generative Models to Improve the Performance of Weak Language Models in Performing Complex Tasks
    US App. 2025/0348745A1
    pending
  • Embedded Attributes for Modifying Behaviors of Generative AI Systems
    US 12,423,338
    granted

    2025

  • Coding activity task (cat) evaluation for source code generators
    US 12,254,293
    granted

    2025

  • Applied machine learning prototypes for hybrid cloud data platform and approaches to developing, personalizing, and implementing the same
    US App. 2023/0267377A1
    pending
  • Detecting Human Input Activity In a Cognitive Environment Using Wearable Inertia and Audio Sensors
    US 11,195,118
    granted

    2021

  • Automated Summarization Based on Physiological Data
    US 10,353,996
    granted

    2019

  • Using Ultraviolet Sensor Data to Determine a Pseudo Location of a User
    US 10,066,987
    granted

    2018

Publications

Full list with filters

39 total

Sort
2026
  • Agents in the Wild: Where Research Meets Deployment
    Grace Hui Yang, Pranav N. Venkit, Hooman Sedghamiz, Enrico Santus, Victor Dibia, Ioana Baldini
    KDD 2026 (Tutorial)
    arXivPDF

    Cite

    Agents in the Wild: Where Research Meets Deployment

    @misc{yang2026agents,
  title = {Agents in the Wild: Where Research Meets Deployment},
  author = {Grace Hui Yang and Pranav N. Venkit and Hooman Sedghamiz and Enrico Santus and Victor Dibia and Ioana Baldini},
  howpublished = {KDD 2026 (Tutorial)},
  year = {2026},
  url = {https://arxiv.org/abs/2607.19336}
}
    tutorial
  • Data Health Evaluation Using Generative Language Models
    Victor Chukwuma Dibia, Chenglong Wang, Bongshin Lee, Jeevana Priya Inala, John Thompson
    Patent (Granted Mar 2026)
    Google Patent

    Cite

    Data Health Evaluation Using Generative Language Models

    @misc{dibia2026data,
  title = {Data Health Evaluation Using Generative Language Models},
  author = {Victor Chukwuma Dibia and Chenglong Wang and Bongshin Lee and Jeevana Priya Inala and John Thompson},
  howpublished = {Patent (Granted Mar 2026)},
  year = {2026},
  url = {https://patents.google.com/patent/US12579115B2/en}
}
    patent
2025
  • Designing Multi-Agent Systems: Principles, Patterns and Implementation for AI Agents
    Victor Dibia
    Book
    Google Books

    Cite

    Designing Multi-Agent Systems: Principles, Patterns and Implementation for AI Agents

    @book{dibia2025designing,
  title = {Designing Multi-Agent Systems: Principles, Patterns and Implementation for AI Agents},
  author = {Victor Dibia},
  year = {2025},
  url = {https://www.google.com/books/edition/Designing_Multi_Agent_Systems/KmyXEQAAQBAJ}
}
    book

    Cited by 6

  • Using Large Generative Models to Improve the Performance of Weak Language Models in Performing Complex Tasks
    Emmanuel Aboah Boateng, Victor Chukwuma Dibia, Cassiano Otavio Becker, Ehimwenma Nosakhare, Nabiha Asghar, Chyna Linn McRae, Anusha Nandam, Omisa Jinsi, Tianwei Chen, Mauricio Cunille Blando, Soundararajan Srinivasan, Damien S Jose, Kabir Walia, Ashwin Srinivasan, Vipul Agarwal, Ananth Rampura Sheshagiri Rao
    Patent
    Justia Patent

    Cite

    Using Large Generative Models to Improve the Performance of Weak Language Models in Performing Complex Tasks

    @misc{boateng2025using,
  title = {Using Large Generative Models to Improve the Performance of Weak Language Models in Performing Complex Tasks},
  author = {Emmanuel Aboah Boateng and Victor Chukwuma Dibia and Cassiano Otavio Becker and Ehimwenma Nosakhare and Nabiha Asghar and Chyna Linn McRae and Anusha Nandam and Omisa Jinsi and Tianwei Chen and Mauricio Cunille Blando and Soundararajan Srinivasan and Damien S Jose and Kabir Walia and Ashwin Srinivasan and Vipul Agarwal and Ananth Rampura Sheshagiri Rao},
  howpublished = {Patent},
  year = {2025},
  url = {https://patents.justia.com/patent/20250348745}
}
    patent

    Cited by 2

  • Magentic-UI: Towards Human-in-the-loop Agentic Systems
    Hussein Mozannar, Gagan Bansal, Cheng Tan, Adam Fourney, Victor Dibia, Jingya Chen, Jack Gerrits, Tyler Payne, Matheus Kunzler Maldaner, Madeleine Grunde-McLaughlin, Eric Zhu, Griffin Bassman, Jacob Alber, Peter Chang, Ricky Loynd, Friederike Niedtner, Ece Kamar, Maya Murad, Rafah Hosn, Saleema Amershi
    arXiv preprint
    arXiv

    Cite

    Magentic-UI: Towards Human-in-the-loop Agentic Systems

    @misc{mozannar2025magentic,
  title = {Magentic-UI: Towards Human-in-the-loop Agentic Systems},
  author = {Hussein Mozannar and Gagan Bansal and Cheng Tan and Adam Fourney and Victor Dibia and Jingya Chen and Jack Gerrits and Tyler Payne and Matheus Kunzler Maldaner and Madeleine Grunde-McLaughlin and Eric Zhu and Griffin Bassman and Jacob Alber and Peter Chang and Ricky Loynd and Friederike Niedtner and Ece Kamar and Maya Murad and Rafah Hosn and Saleema Amershi},
  howpublished = {arXiv preprint},
  year = {2025},
  url = {https://arxiv.org/html/2507.22358v1}
}
    preprint

    Cited by 64

  • Concept Distillation from Strong to Weak Models via Hypotheses-to-Theories Prompting
    Emmanuel Aboah Boateng, Cassiano O. Becker, Nabiha Asghar, Kabir Walia, Ashwin Srinivasan, Ehi Nosakhare, Victor Dibia, Soundar Srinivasan
    NAACL 2025
    arXivPDF

    Cite

    Concept Distillation from Strong to Weak Models via Hypotheses-to-Theories Prompting

    @inproceedings{boateng2025concept,
  title = {Concept Distillation from Strong to Weak Models via Hypotheses-to-Theories Prompting},
  author = {Emmanuel Aboah Boateng and Cassiano O. Becker and Nabiha Asghar and Kabir Walia and Ashwin Srinivasan and Ehi Nosakhare and Victor Dibia and Soundar Srinivasan},
  booktitle = {NAACL 2025},
  year = {2025},
  url = {https://arxiv.org/abs/2408.09365}
}
    conference

    Cited by 4

  • Interactive Debugging and Steering of Multi-Agent AI Systems
    Will Epperson, Gagan Bansal, Victor Dibia, Adam Fourney, Jack Gerrits, Erkang Zhu, Saleema Amershi
    CHI 2025, Yokohama, Japan

    Cite

    Interactive Debugging and Steering of Multi-Agent AI Systems

    @inproceedings{epperson2025interactive,
  title = {Interactive Debugging and Steering of Multi-Agent AI Systems},
  author = {Will Epperson and Gagan Bansal and Victor Dibia and Adam Fourney and Jack Gerrits and Erkang Zhu and Saleema Amershi},
  booktitle = {CHI 2025},
  year = {2025},
  url = {https://chi2025.acm.org/}
}
    conference

    Cited by 125

  • AI Fatigue: Reflections on the Human Side of AI's Rapid Advancement
    Victor Dibia
    Communications of the ACM
    DOI

    Cite

    AI Fatigue: Reflections on the Human Side of AI's Rapid Advancement

    @article{dibia2025fatigue,
  title = {AI Fatigue: Reflections on the Human Side of AI's Rapid Advancement},
  author = {Victor Dibia},
  journal = {Communications of the ACM},
  year = {2025},
  url = {https://doi.org/10.1145/3734515}
}
    journal

    Cited by 3

2024
  • Challenges in Human-Agent Communication
    Gagan Bansal, Jennifer Wortman Vaughan, Saleema Amershi, Eric Horvitz, Adam Fourney, Hussein Mozannar, Victor Dibia, Daniel S. Weld
    Microsoft Research Technical Report
    PDF

    Cite

    Challenges in Human-Agent Communication

    @techreport{bansal2024challenges,
  title = {Challenges in Human-Agent Communication},
  author = {Gagan Bansal and Jennifer Wortman Vaughan and Saleema Amershi and Eric Horvitz and Adam Fourney and Hussein Mozannar and Victor Dibia and Daniel S. Weld},
  institution = {Microsoft Research Technical Report},
  year = {2024},
  url = {https://www.microsoft.com/en-us/research/publication/human-agent-interaction-challenges/}
}
    technical report

    Cited by 73

  • Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks
    Adam Fourney, Gagan Bansal, Hussein Mozannar, Cheng Tan, Eduardo Salinas, Erkang (Eric) Zhu, Friederike Niedtner, Grace Proebsting, Griffin Bassman, Jack Gerrits, Jacob Alber, Peter Chang, Ricky Loynd, Robert West, Victor Dibia, Ahmed Awadallah, Ece Kamar, Rafah Hosn, Saleema Amershi
    Microsoft Research Technical Report
    Project PagePDF

    Cite

    Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks

    @techreport{fourney2024magentic,
  title = {Magentic-One: A Generalist Multi-Agent System for Solving Complex Tasks},
  author = {Adam Fourney and Gagan Bansal and Hussein Mozannar and Cheng Tan and Eduardo Salinas and Erkang (Eric) Zhu and Friederike Niedtner and Grace Proebsting and Griffin Bassman and Jack Gerrits and Jacob Alber and Peter Chang and Ricky Loynd and Robert West and Victor Dibia and Ahmed Awadallah and Ece Kamar and Rafah Hosn and Saleema Amershi},
  institution = {Microsoft Research Technical Report},
  year = {2024},
  url = {https://aka.ms/magentic-one}
}
    technical report

    Cited by 235

  • Data Analysis in the Era of Generative AI
    Jeevana Priya Inala, Chenglong Wang, Steven Drucker, Gonzalo Ramos, Victor Dibia, Nathalie Riche, Dave Brown, Dan Marshall, Jianfeng Gao
    arXiv preprint
    arXivPDF

    Cite

    Data Analysis in the Era of Generative AI

    @misc{inala2024data,
  title = {Data Analysis in the Era of Generative AI},
  author = {Jeevana Priya Inala and Chenglong Wang and Steven Drucker and Gonzalo Ramos and Victor Dibia and Nathalie Riche and Dave Brown and Dan Marshall and Jianfeng Gao},
  howpublished = {arXiv preprint},
  year = {2024},
  url = {https://arxiv.org/abs/2409.18475}
}
    preprint

    Cited by 28

  • AutoGen Studio: A No-Code Developer Tool for Building and Debugging Multi-Agent Systems
    Victor Dibia, Jingya Chen, Gagan Bansal, Suff Syed, Adam Fourney, Erkang Zhu, Chi Wang, Saleema Amershi
    EMNLP 2024, arXiv.org
    PDFCode

    Cite

    AutoGen Studio: A No-Code Developer Tool for Building and Debugging Multi-Agent Systems

    @inproceedings{dibia2024autogen,
  title = {AutoGen Studio: A No-Code Developer Tool for Building and Debugging Multi-Agent Systems},
  author = {Victor Dibia and Jingya Chen and Gagan Bansal and Suff Syed and Adam Fourney and Erkang Zhu and Chi Wang and Saleema Amershi},
  booktitle = {EMNLP 2024},
  year = {2024},
  url = {https://arxiv.org/abs/2408.15247}
}
    conference

    Cited by 63

  • Towards better Human-Agent Alignment: Assessing Task Utility in LLM-Powered Applications
    Negar Arabzadeh, Julia Kiseleva, Qingyun Wu, Chi Wang, Ahmed Awadallah, Victor Dibia, Adam Fourney, Charles Clarke
    arXiv preprint
    PDF

    Cite

    Towards better Human-Agent Alignment: Assessing Task Utility in LLM-Powered Applications

    @misc{arabzadeh2024towards,
  title = {Towards better Human-Agent Alignment: Assessing Task Utility in LLM-Powered Applications},
  author = {Negar Arabzadeh and Julia Kiseleva and Qingyun Wu and Chi Wang and Ahmed Awadallah and Victor Dibia and Adam Fourney and Charles Clarke},
  howpublished = {arXiv preprint},
  year = {2024},
  url = {https://arxiv.org/abs/2402.09015}
}
    preprint

    Cited by 13

  • Coding activity task (cat) evaluation for source code generators
    Victor Dibia, Adam Fourney, Forough POURSABZI SANGDEH, Saleema Amin Amershi
    Patent (Granted Mar 2025)
    Google Patent

    Cite

    Coding activity task (cat) evaluation for source code generators

    @misc{dibia2024coding,
  title = {Coding activity task (cat) evaluation for source code generators},
  author = {Victor Dibia and Adam Fourney and Forough POURSABZI SANGDEH and Saleema Amin Amershi},
  howpublished = {Patent (Granted Mar 2025)},
  year = {2024},
  url = {https://patents.google.com/patent/US12254293B2/en}
}
    patent

    Cited by 3

2023
  • Axiomatic Preference Modeling for Longform Question Answering
    Corby Rosset, Guoqing Zheng, Victor Dibia, Ahmed Awadallah, Paul Bennett
    EMNLP 2023, Singapore, EMNLP 2023
    PDFModel

    Cite

    Axiomatic Preference Modeling for Longform Question Answering

    @inproceedings{rosset2023axiomatic,
  title = {Axiomatic Preference Modeling for Longform Question Answering},
  author = {Corby Rosset and Guoqing Zheng and Victor Dibia and Ahmed Awadallah and Paul Bennett},
  booktitle = {EMNLP 2023},
  year = {2023},
  url = {https://arxiv.org/abs/2312.02206}
}
    conference

    Cited by 8

15 of 39 shown

Talks

12 total

2025
2024
2023
2021
2019

Read the Newsletter.

I write a monthly newsletter on Applied AI and HCI. Subscribe to get notified on new posts.

Feel free to reach out! Twitter, GitHub, LinkedIn

Read and Subscribe