Dheeraj Parkash

Intro

I'm a Data Science and Applied AI practitioner with a strong academic and hands-on background in machine learning, NLP, and predictive modeling. Currently completing my 𝐌𝐚𝐬𝐭𝐞𝐫'𝐬 𝐚𝐭 𝐔𝐧𝐢𝐯𝐞𝐫𝐬𝐢𝐭é 𝐏𝐚𝐫𝐢𝐬-𝐒𝐚𝐜𝐥𝐚𝐲 𝐚𝐧𝐝 𝐄𝐍𝐒 𝐏𝐚𝐫𝐢𝐬-𝐒𝐚𝐜𝐥𝐚𝐲 (𝐌𝐕𝐀 𝐭𝐫𝐚𝐜𝐤), I focus on building scalable, interpretable, and high-performing models that address complex real-world challenges. My experience spans both applied research and industry innovation. At 𝑨𝑳𝑻𝑬𝑵, I work on emotion recognition models using multimodal Transformer architectures to personalize VR interactions. Previously at 𝑰𝑵𝑹𝑰𝑨/𝑾𝑰𝑴𝑴𝑰𝑪𝑺, I contributed to 𝒉𝒂𝒕𝒆 𝒔𝒑𝒆𝒆𝒄𝒉 𝒅𝒆𝒕𝒆𝒄𝒕𝒊𝒐𝒏 𝒔𝒚𝒔𝒕𝒆𝒎𝒔 for multi-party conversations, combining state-of-the-art text and graph methods to enhance model accuracy and social context understanding. I enjoy working at the intersection of research and application—leveraging tools like Python, PyTorch, Transformers, and LLMs to turn abstract ideas into deployable AI systems. My recent work includes projects on recommender systems, social bias detection, and sports prediction, with an emphasis on performance optimization, explainability (𝑿𝑨𝑰), and error analysis. Curiosity, cross-functional collaboration, and a drive to create meaningful impact guide my work. 𝑳𝒆𝒕’𝒔 𝒄𝒐𝒏𝒏𝒆𝒄𝒕 𝒂𝒏𝒅 𝒆𝒙𝒄𝒉𝒂𝒏𝒈𝒆 𝒊𝒅𝒆𝒂𝒔 𝒂𝒓𝒐𝒖𝒏𝒅 𝒂𝒑𝒑𝒍𝒊𝒆𝒅 𝑨𝑰, 𝒊𝒏𝒕𝒆𝒍𝒍𝒊𝒈𝒆𝒏𝒕 𝒔𝒚𝒔𝒕𝒆𝒎𝒔, 𝒂𝒏𝒅 𝒔𝒐𝒍𝒗𝒊𝒏𝒈 𝒎𝒆𝒂𝒏𝒊𝒏𝒈𝒇𝒖𝒍 𝒄𝒉𝒂𝒍𝒍𝒆𝒏𝒈𝒆𝒔 𝒘𝒊𝒕𝒉 𝒅𝒂𝒕𝒂.

Links