"With your eyes on the sky and your feet on the ground, chase your dream relentlessly."
Christos Petridis
PhD student in CIS @ Temple
Christos Petridis
PhD student in CIS @ Temple
About Me
I am a 3rd year PhD student at Temple University in Philadelphia, PA. My PhD is in Computer and Information Sciences and my advisor is Prof. Zoran Obradovic.
📌 My research focuses on whether Large Language Models (LLMs) can interpret real-world environmental data, such as weather conditions, well enough to serve as reliable components in predictive pipelines compared to machine learning. In the same area, I assess how LLMs perceive verbal probability expressions such as "possibly" and "unlikely" compared to human baselines. On the data management side, I work with Prof. Eduard Dragut on dataset cartography, making the diagnostics from data maps cheap enough to be practical at scale.
📌 Before my Ph.D., I got an Integrated Masters (5 years) in Electrical and Computer Engineering (ECE) from the University of Thessaly in Volos, Greece. I completed my thesis (and published a paper) in collaboration with Angelicoussis Group where I worked on estimating hull fouling* using machine learning and propulsion data. During my studies, I also interned at Angelicoussis Group (thesis collaboration 2023) and other software companies.
📌 I've been fortunate to collaborate with Prof. Konstantinos Pelechrinis (Pitt) and Prof. Mladen Kezunovic (Texas A&M).
Currently, I am building Carlytics.gr
* Hull Fouling is the undesirable accumulation of marine organisms on submerged structures, increasing drag and fuel use.
Education
Ph.D. in Computer and Information Sciences
Advisor: Dr. Zoran Obradovic
GPA: 3.93/4.00
5-year Integrated Masters in Electrical and Computer Engineering (300 ECTS)
Thesis: "Detecting Hull Fouling using Machine Learning Algorithms trained on Ship Propulsion Data", advised by Dr. Michael Vassilakopoulos
GPA: 8.23/10.0 (Ranked 4th in my class, Top 10% of the academic year)
Professional Experience
Graduate Research & Teaching Assistant
Temple University
- •UAV Navigation: Developing methods to predict the location and eventually the trajectory of UAVs (drones) in GPS-denied environments using visual features extracted from aerial imagery.
- •Power Outage Prediction: Working with spatiotemporal data to predict weather-related power outages and their duration.
- •TA for CIS 2109 'Database Management Systems' (Fall 2026) Instructor: Dr. Nancy Polychronopoulou
CTO & Co-Founder
Carlytics
- •Building the technological infrastructure (including the backend API service, AWS cloud architecture, and database administration) for the smartest way to analyze used cars and interact with data through natural language, enabling comprehensive, data-driven insights and confident predictions for the (used) car market in Greece.
Data Science Intern
Angelicoussis Group
- •Developed a framework to extract text from company's documents and classify them into categories using NLP tecnhiques (fine tuned pre-trained language models like DistilBERT, XLNet etc.).
- •Performed fleet performance prediction & evaluation using data related to company's ships.
- •Performed exploratory data analysis to select the best anti-fouling hull paint examining different factors for the whole fleet.
- •Worked with the R&D team on estimation of added resistance for vessels, aiming to reduce environmental impact and improve fuel management (Integrated Master's thesis collaboration).
Software Engineer Intern
DevN (Psathas Neilos Christos Software Company)
- •Assisted in development of the front-end of two mobile (Android) applications using Java.
- •Collaborated with team members using version control systems such as Git to organize modifications and updates.
- •Worked with Google Firebase to manage user inputted data across the mobile applications.
- •Assisted in the backend development of a web app (API service) utilizing Node.js and Express.js.
Software Engineer Intern
Swollet Technologies Ltd.
- •Assisted in development of the front-end of two mobile (Android) applications using Java.
- •Collaborated with team members using version control systems such as Git to organize modifications and updates.
- •Worked with Google Firebase to manage user inputted data across the mobile applications.
- •Assisted in the backend development of a web app (API service) utilizing Node.js and Express.js.
Research Papers (first author)
Papers are presented in chronological order (with the most recent appearing first).
Evaluating Large Language Models for Forced Outage Risk Prediction: Benefits and Comparison to Machine Learning
Authors: Christos Petridis, Zoran Obradovic, Mladen Kezunovic
Venue: (in press) 60th Hawaii International Conference on System Sciences, (HICSS 2027)
How Unlikely Is "Unlikely"? Assessing Verbal Probability Perception Across Large Language Models
Authors: Christos Petridis, Konstantinos Pelechrinis, Zoran Obradovic
Venue: arXiv
From Prior Beliefs to Lineup Truths: Bayesian Inference for Lineup Performance
Authors: Christos Petridis, Konstantinos Pelechrinis, Zoran Obradovic
Venue: under review (preprint from Research Square)
A Cost-Aware Evaluation of Duration Predictions for Weather-Induced Forced Power Outages
Authors: Christos Petridis, Zoran Obradovic, Rashid Baembitov, Mladen Kezunovic
Venue: 22nd International Conference on Artificial Intelligence Applications and Innovations (AIAI 2026)
Lineup Regularized Adjusted Plus-Minus (L-RAPM): Basketball Lineup Ratings with Informed Priors
Authors: Christos Petridis, Konstantinos Pelechrinis
Venue: arXiv
PixelPath: Predicting UAV Trajectories in GPS-Restricted Environments Using Image Feature Extraction and Machine Learning
Authors: Christos Petridis, Abhudaya Shrivastava, Marijana Vacic, Zoran Obradovic
Venue: 21st International Conference on Artificial Intelligence Applications and Innovations (AIAI 2025)
Detecting Hull Fouling using Machine Learning Algorithms trained on Ship Propulsion Data to Improve Resource Management and Increase Environmental Benefits
Won the Best Paper award in Smart Green category
Authors: Christos Petridis, Michael Vassilakopoulos
Venue: 8th International Conference on Smart Data and Smart Cities (SDSC 2024)
Teaching Experience
CIS 2109 Database Management Systems
ECE 311 Database Systems I
ECE 326 Object Oriented Programming
News & Updates
Paper accepted at HICSS
Aug 2026
PublicationExcited to share that our paper "Evaluating Large Language Models for Forced Outage Risk Prediction: Benefits and Comparison to Machine Learning" has been accepted at the 60th Hawaii International Conference on System Sciences (HICSS).
Attended the North East AI Agents Day 2026 @ Jane Street's Headquaters in NYC
May 2026
ConferenceThe goal of this workshop is to offer a comprehensive overview of AI agents, bring ML, Systems, and HCI research communities together to share progress, discuss common problems and evaluation setups, and identify opportunities for collaboration.
Attended the North East Database Day 2026 @ UMass Boston
January 2026
ConferenceThe North East Database Day (NEDB Day) is an annual one-day academic and industry conference focused on database systems, data management, analytics, and related areas of data-intensive computing.
Successfully passed my PhD Qualifying Exam
January 2026
AchievementThe Qualifying Examination tests the student on the fundamentals of Computer and Information Science and the knowledge required to do research in the field. It consists of a written exam on theory and algorithms, systems, and track-specific material.
Virtually presented three papers at AIAI 2025
June 2025
Conference- •PixelPath: Predicting UAV Trajectories in GPS-Restricted Environments Using Image Feature Extraction and Machine Learning
- •Spatiotemporal Multiplex Network Model for Predicting Forced Outage Severity in Distribution Grids
- •Autonomous Navigation in Swarm of UAVs Using Spatio Temporal Data and Constrained-Reinforcement Learning
Best Paper Award at SDSC 2024
July 2024
AwardOur paper entitled 'Detecting Hull Fouling using Machine Learning Algorithms trained on Ship Propulsion Data to Improve Resource Management and Increase Environmental Benefits' won the Best Paper Award in the Smart Green category at the 8th International Conference on Smart Data and Smart Cities (SDSC 2024).