Lifelong learner.
I have been working on deep neural networks for the past few years, while serving as an applied researcher. On the heels of built experience, I have returned to fundamental/theoretical research.
Research interests include generative modeling (discrete, continuous, with diffusion, flows, stochastic interpolants, posterior sampling etc), representation learning and multimodality, employing transformers and other neural network architectures, graphs, signals, and the frequency domain. Having trained in both the theoretical and the computational domains, I am interested in interesting/open problems related to artificial intelligence, including the building blocks and fundamental aspects of deep learning, such as theoretical understanding, model design, and learning algorithms.
​
Have published in relevant conferences and journals (eg. ICML'24 AI4MATH workshop, NeurIPS'24 Multimodal Algorithmic Reasoning workshop, ECIR, SIGIR-AP'25, SIAM's Algo Diff.'24, etc.). TensorFlow, MXNet, and HuggingFace Transformers open source contributor of deep learning research.
​
My PhD research (mathematical statistics, Virginia Tech) had been in algorithms related to maximum likelihood estimation, KL divergence, non-Gaussian processes, and control (research funded by private grants). During my MSc at VT, I had the honor to briefly collaborate with I.J.Good.
Majored in computer science and applied cybernetics/math (N. Wiener style) for my undergraduate degree, when I was first introduced to artificial neural networks. For good measure, I took Z. Kira's Deep Learning graduate course at Georgia Tech CS.
Theoretic highschool -CS track. Math and physics (thermodynamics) olympic (silver medal, Romania nationals).
​
Research highlights on X(Twitter) during my tenure as Staff ML Engineer / Applied Researcher@ NYC Tech Industry (ex- ASAPP, Amazon, Reddit, Etsy):
* Designed and trained a (deep) neural network architecture for representation learning in personalized sponsored search and recommendations
*Advocated for more DL theory research (2017)
​
Dabbled in algorithmic trading for a nat (CFA charterholder).
US and EU dual citizen. Located in NYC.
​
​
​
​
​
​
​
​