About me
I am a 5th-year Ph.D. candidate at the University of California, Irvine, working under the mentorship of Professor Pierre Baldi with a primary focus on Deep Learning. My research centers on neural architectures—particularly attention mechanisms, transformers, and bio-inspired methods—for speech and multimodal applications, including robust LLMs and speech-enabled LLMs. Alongside my technical strengths, I pride myself on clear communication, collaborative teamwork, and a solid work ethic, all of which drive me to excel in demanding research environments.
My journey began at the Electrical and Computer Engineering Department at the University of Thessaly(UTH), where I graduated 4th in my class with a minor in Data Science. Under the guidance of Gerasimos Potamianos, I completed my undergraduate thesis on Sign Language Video-to-Text Translation. During my time there, I founded HERMES—a powered exoskeleton for Cybathlon 2020—along with the UTH Trading and Investing Club, and served as both Chair and Vice Chair of the IEEE SB Volos. I have also gained practical industry exposure at Hilti (Switzerland), participated in the Oactive project at CERTH (Volos), contributed remotely to In4Capital (London), and conducted research with SML Lab (Cyprus) under Sotirios Chatzis.
In my free time, I’m typically either delving into a good book or endurance training—gym workouts and trail running both keep me balanced and motivated.
Skills
Languages / Libraries
Python, PyTorch, TensorFlow, NumPy, Pandas, HuggingFace, PyTorch Lightning
Machine Learning Expertise
Deep Learning, Generative Models (TTS, Diffusion), Transformer Architectures, Attention Mechanisms, Speech Research, Speech Editing, Adversarial Training, Bayesian NNs, Speech Language Models, LLMs
Tools & Platforms
AWS (S3, EC2, SageMaker), Git, Unix/Linux, Jupyter, LaTeX
Data Science & Analytics
Data Preprocessing, Feature Engineering, Statistical Analysis, Experiment Design, Visualization (Matplotlib, Seaborn), Agile Development
News
- December 2024 Our work Improving Deep Learning Speed and Performance through Synaptic Neural Balance has been accepted to AAAI 2025.
- October 2024 Our work Improving Deep Learning Speed and Performance through Synaptic Neural Balance has been accepted to 4 workshops in NeurIPS 2024.
- October 2024 Our work Domain-Adaptive ML for Surface Roughness Predictions in Nuclear Fusion has been accepted to NeurIPS 2024 Workshop on Machine Learning and the Physical Sciences.
- October 2024 Our work Cold Posterior Effect towards Adversarial Robustness has been accepted to NeurIPS 2024 Workshop on Bayesian Decision-making and Uncertainty.
- September 2024 Our work Nuclear Fusion Diamond Polishing Dataset has been accepted as a Dataset paper in NeurIPS 2024.
- June 2024 Our work Precision polishing of ablator capsules via in-situ process monitoring and machine-learning-based optimization with LLNL has been accepted to Fusion Science and Technology journal.
- February 2024 I will join AWS Transcribe team to work on the safety alignment with adversarial defenses of one of the first Speech Language Models.
- December 2023 Our work Machine Learning-Enhanced Prediction of Surface Smoothness for Inertial Confinement Fusion Target Polishing Using Limited Data with LLNL has been accepted to AIM 2024 as an Extended Abstract.
- November 2023 Our paper FastStitch: Towards speech editing by hitch-hiking a pre-trained FastSpeech2 model has been accepted to NLDL 2024 as an oral.
- October 2023 Our paper AttentionStitch: How Attention Solves the Speech Editing Problem has been accepted to NeurIPS 2023 Workshop on Machine Learning for Audio.
- May 2022 Our paper “Structured Stochastic Gradient MCMC: a hybrid VI and MCMC approach” has been accepted to ICML 2022.
- March 2022 I will join Disney Research at Walt Disney Imagineering this summer for a research internship on Text-to-Speech models for Speech Editing.
- November 2021 Our paper “Structured Stochastic Gradient MCMC: a hybrid VI and MCMC approach” has been accepted as a spotlight talk in NeurIPS Bayesian Deep Learning workshop 2021 and as a poster in AABI 2022. You can find the video presentation here.
- January 2021 Our paper “Local Competition and Stochasticity for Adversarial Robustness in Deep Learning” has been accepted to AISTATS 2021.
- August 2020 I successfully defended my undergraduate thesis on Sign Language Translation which you can find here.
- June 2020 Our paper “Prediction of pain in knee osteoarthritis patients using machine learning: Data from Osteoarthritis Initiative” with CERTH has been accepted to IISA20.
- May 2020 In September 2020 I will join UCI as a PhD student.
- February 2020
- Our paper “Physical activity as a risk factor in the progression of Osteoarthritis: A machine learning perspective” has been accepted to LION14.
- In March I will join HILTI in Switzerland for a data science internship.
- October 2019 Our paper “Which attributes matter the most for loan origination? A neural attention approach.” has been accepted to Robust AI for Financial Services Workshop in NeurIPS 2019.