I am a passionate AI PhD student and engineer focused on deep multimodal models, generative frameworks, and robotics intelligence.

My experience spans deep model implementation and training, under-the-hood knowledge and research, as well as pipeline integration. My research involves applications of reinforcement learning for the next generation of language models.

Please feel free to view my recent work below.
AI research and development for NLP, decision making, and robotics.
Research
Diffusion LLMs, Bandits
Reinforcement Learning for Diffusion LLMs
Investigating the use of reinforcement learning during latent sampling process for reasoning and agentic diffusion language models.
Attacks on Stochastic Dueling Bandits
We provide a mathematical framework for reward poisoning on stochastic dueling bandits algorithms, proving effective spoofing with sublinear cost bounds and forced linear regret. Paper available on OSU Archives, code available on Github.
Projects
NLP, Generation, Translation
Thompson Sampling with Diffusion and Flow Matching Priors for Linear Banditsl
Developed guided diffusion and flow matching algorithms for linear bandit prior sampling, as well as low-rank approximations to reduce per-round time complexity while preserving regret bounds. Open-source code available on Github.
Discrete Diffusion and Auto-Regressive Language Model
A from-scratch implementation of both traditional auto-regressive and discrete diffusion generative transformer-based language models. Trained on openwebtext utilizing DGX system. Open-source code available on Github.
Denoising Diffusion and Flow Matching Image Generation
A from-scratch implementation of denoising diffusion and flow matching image generative models (CNN U-Net and latent vision transformer). Trained on Stanford Cars, CelebA, ImageNet utilizing DGX system. Open-source code available on Github.
Experience
Industry, Educational
REVOBOTS AI Intern
Designed and deployed video conferencing application allowing real-time VR embodiment of robots for teleoperation. Implemented with WebXR, GPU accelerated encoding, SFU server.

Utilized Hugging Face LeRobot framework for teleoperation, 100+ episode dataset recording, compression.

Employed MLOps for automated multi-day training of transformer-based robot policies on DGX system using collected datasets.
Graduate Teaching Assistant
Teaching assistant for Programming Language Fundamentals, covering Haskell and Prolog.
Undergraduate Learning Assistant
Learning assistant for Operating Systems I, Data Structures, and Gen Engineering.