cv

Carnegie Mellon University – Ph.D. in Machine Learning.

• Awarded the CMU Presidential Fellowship.
• GPA: 4.18 (4.0 scale).
• Advisor: Prof. Tom M. Mitchell.
• I worked on the Never Ending Language Learning (NELL) project.
• My thesis work was on Learning Collections of Functions.
• I performed research on topics related to self-reflection in machine learning:
  • I developed multiple methods for using unlabeled data to estimate the accuracies of several different classifiers performing the same task, that I have presented at Uncertainty in Artificial Intelligence (UAI) in 2014, International Conference in Machine Learning (ICML) in 2016, and at Neural Information Processing Systems (NIPS) in 2017.
  • I developed a method for performing low-resource and zero-shot multilingual machine translation.
  • I am developing a self-reflection mechanism for machine learning systems and I am applying it in NELL.
  • I am working on a new framework for learning where agreement among multiple agents is the objective.

Carnegie Mellon University – M.S. in Machine Learning.

• GPA: 4.18 (4.0 scale).
• Advisor: Prof. Tom M. Mitchell.
• Thesis: Estimating Accuracy from Unlabeled Data.
• Master's degree requirements completed while working towards obtaining my Ph.D. in machine learning.

Imperial College London – M.Eng. and ACGI in Electrical and Electronic Engineering.

• Integrated Bachelor's and Master's degree.
• Dean’s List for exceptional academic performance (all four years).
• Grade: 91.67% (top 1% of class) with 92.37% on the final exams – First-Class Honors Degree.
  • Equivalent to 4.0 GPA in the USA academic system.
• Thesis Title: Human Motion Classification Using Statistical Machine Learning Methods.
  • Awarded the Sir Bruce White prize in engineering for the best final year project.
• Third Year Summer Research Project: "Bayesian Machine Learning Methodologies for Modeling Time Series with High Volatility".
  • Volatility modeling and prediction in a multitude of applications with significance both for industry (e.g. finance) and academia.

Research Scientist at Scaled Cognition.

• Developing a new generation of rational, controllable AI models deployable as domain experts for grounded, real-world applications.

Senior Principal Researcher (L67) at Microsoft Semantic Machines.

• Served as a technical leader, managing a team across both research and engineering, and helping set technical direction for the broader organization. I invested heavily in aligning research and engineering, and I worked closely with multiple partner teams at Microsoft to deliver on our objectives, while also helping improve the underlying platform and infrastructure that we leveraged.

Principal Researcher (L65) at Microsoft Semantic Machines.

• Led the design and development of a system that learns to predict the next user action (e.g., sending an email, opening a file, adding specific people to a calendar invite, etc.) based on what a user and others in their network have been doing. Underlying this system is a novel Foundation Model over user actions.

Senior Researcher (L64) at Microsoft Semantic Machines.

• Manager: Dan Klein, Technical Fellow.
• Designed and shipped a contextual semantic parser that powers the Semantic Machines conversational AI platform.
• Worked on multiple research projects leading to academic publications at the intersection of natural language processing and machine learning.

Research Analyst Consultant at Cubist Systematic Strategies.

• Worked for the Advanced Methods Groups (AMG) in new machine learning-driven approaches for end-to-end algorithmic trading systems.
• Developed the main pipeline for trading algorithms research and for integrating these algorithms in a new production system.

Research Intern at Microsoft Research.

• Manager: Dr. Eric Horvitz, Distinguished Scientist and Managing Director of Microsoft Research.
• Collaborators: Dr. Hoifung Poon and Dr. Ashish Kapoor.
• First 3 months: Developed a probabilistic logic-based approach for accuracy estimation using unlabeled data, as part of the Literome Project.
• Last 3 months: Active learning amidst logical knowledge.

Co-Founder, President and Chief Technology Officer (CTO) at Holic.

• Developed an intelligent news reader application.
• Developed algorithms utilizing advanced nonparametric Bayesian models for clustering news articles and for user profiling.
• Received €400,000 funding from the following investors:
  • Dr. John Coustas, President & CEO of Danaos Corporation.
  • Mr. Leon Yohai, Founder & CEO of ZuluTrade.

Founder of Project Protasis – Forest Protection.

• Developed a forest fire detection and prevention system.
• Developed both hardware and software (involving use of neural networks for recognizing fire and smoke in photos).
• Installed prototype in “Alsos Syggrou”, a wooded area in Athens, funded by the Greek government.
• Got patent approved in Greece (Patent #: 20110100016) and international patent pending approval.
• Interviewed by CNN, as well as several newspapers, magazines and other TV channels, regarding this project.
• Won the 2^nd place in the Microsoft Imagine Cup 2009 Greek Finals with this project.

Web Designer and Software Developer at Curtainmakers S.A.

Founder of Project Protasis – Home Automation.

• Developed a smart home automation system that controls all electrical appliances, heating systems and water systems and aims to reduce domestic energy consumption while making the lives of homeowners as comfortable as possible.
• Developed software, employing several machine learning algorithms.
• Was funded by the Greek government and more specifically, by the Organization of School Buildings.
• Interviewed by NBC, as well as several newspapers, magazines and other TV channels, regarding this project.
• Won the Microsoft Imagine Cup 2008 Greek Finals with this project and represented Greece in the Imagine Cup 2008 Worldwide Finals, in Paris, France, competing against teams from universities, at age of only 16 and having worked without a team or a mentor.

Speculative Execution of Dataflow Program Nodes (US 17/815,904).

Semantic Parsing of Utterance Using Contractive Paraphrasing (US 17/229,637).

TensorFlow Scala.

• Machine learning library and API for the TensorFlow framework developed by Google, that is currently being used by multiple researchers and companies (~900 GitHub stars).

Swift for TensorFlow.

• I was a contributor tp this project, which is a machine learning library and API for the TensorFlow framework developed by Google, and that enables differentiable programming.

Symphony MT.

• Machine translation library written in Scala, that supports various models along the whole translation pipeline (i.e., ranging from data preprocessing and vocabulary generation, to multiple encoder and decoder models).

Makina.

• Machine learning and optimization library written in Java that includes a scalable implementation of the Probabilistic Soft Logic (PSL) framework.

• 2^nd place in the CMU NeuroHackathon.

• CMU Presidential Fellowship.

• Sir Bruce White prize in engineering for the best M.Eng. thesis at Imperial College London.
• Dean's List at Imperial College London, for exceptional academic performance (all four years).

• Nominated for the RAE (UK Royal Academy of Engineering) presentation skills award.
• "Roll of Honor" of the Electrical and Electronic Engineering department of Imperial College London, for the best software design project in the year.

• Congratulatory diploma for my work on "Project Protasis – Home Automation" by UNESCO.
• Honorary plaque by the Greek Minister of Education and Religious Affairs for honoring my country internationally.

• Full scholarship (of about €12,000) by "Geitonas School", for honoring my school internationally.
• Had a computer science laboratory named after me in “Geitonas School”, for honoring my school internationally.
• Awarded €1,000 by the President of the Greek Parliament, for honoring my country internationally.
• Honorary plaque by the Mayor of my city, Vari, in Greece, for honoring my country internationally.
• Represented Greece in the Microsoft Imagine Cup 2008 Worldwide Finals, in Paris, France, with "Project Protasis - Home Automation".
• 1^st place in the Microsoft Imagine Cup 2008 Greek Finals with "Project Protasis - Home Automation", competing against teams from universities, at age of only 16 and having worked without a team or a mentor (also won the 2^nd place in the Microsoft Imagine Cup 2009 Greek Finals with "Project Protasis - Forest Protection").

• Certificate of Excellence for the school year 2006-07 from the Greek Ministry of Education and Religious Affairs.
• 3^rd place (bronze medal) in the Greek Mathematics Olympiad, organized by the Hellenic Mathematical Society.

• 11 scholarships and competitions in the fields of mathematics, sciences and computer science.

Provided support for an advanced graduate-level course on machine learning at Carnegie Mellon University.

• Class: 10-812 Architectures for Never-Ending Learning.
• Taught by Prof. Tom Mitchell.
• Gave a couple of lectures and helped with the organization of the class.
• Designed and co-led the implementation of the software testbed used to support the course projects.
• Mentored several students working on class projects.

Teaching Assistant for a graduate-level machine learning course at Carnegie Mellon University.

• Class: 10-701/15-781 Introduction to Machine Learning.
• Did this for two semesters (taught by Prof. Geoff Gordon and Prof. Aarti Singh in 2014, and Prof. Tom Mitchell in 2016).
• Gave some lectures and some recitation lectures.
• Mentored several groups of students working on class projects.
• Wrote and graded homework assignments and exams.

Competence-based Curriculum Learning for Neural Machine Translation.

• North American Chapter of the Association for Computational Linguistics (NAACL) conference.

TensorFlow Scala.

• Invited talk at Google.

Estimating Accuracy from Unlabeled Data.

• Neural Information Processing Systems (NIPS) conference.
• International Conference in Machine Learning (ICML) conference.
• Uncertainty in Artificial Intelligence (UAI) conference.
• Invited talk at Google.
• AI Lunch at Stanford University.
• ML Lunch at Carnegie Mellon University.

Nonparametric Mixtures of Multi-Output Heteroscedastic Gaussian Processes.

• "Nonparametric Methods in Machine Learning" workshop at NIPS.

Project Protasis - Forest Protection.

• Microsoft Hellas Headquarters.
• Ministry of Transportation of the Greek government (the Ministry later funded a pilot program).

Project Protasis - Home Automation.

• Microsoft Hellas Headquarters.
• Louvre Museum in Paris, France, as part of Imagine Cup.
• UNESCO International Conference on Climate Change and Challenges for the Future Generations.
• Organization of School Buildings of the Greek government (later received funding for a pilot program from the Ministry of Education and Religious Affairs).
• 73^rd Thessaloniki International Fair.
• Generation Next 0-18 fair, in Athens.
• 4^th International Conference on Information and Communication Technologies in Bio and Earth Sciences (HAICTA 2008), in Athens.