Hi, I am Suraj Parmar, currently working as a Machine Learning Engineer at Maisa AI.
Previously, I worked at a medical-legal start-up in Toronto where I managed the cloud infrastructure and built scalable agentic document processing pipelines handling hundreds of customers daily. Before that, I worked at SpassMed, a healthcare start-up in Toronto, where I developed statistical and deep learning based forecasting and classification models, built efficient Retrieval Augmented Generation (RAG) pipelines for offline environments, and published papers on early Sepsis detection.
My master's thesis was on the topic of "Utilizing Deep Learning for Enhancing Performance on Encrypted Stock Market Data". The goal was to extract robust and meaningful representations for Numerai datasets to improve the GBDT baseline model. I have been researching on the Numerai datasets for financial data. My main focus right now is Transformers on cross-sectional asset prediction involving self-supervised learning. Can read more about it on my blog.
Earlier, I interned at Zujo, a Computer Vision start-up in India, and later worked at RealAdvisor as a Data Scientist remotely.
I am available for contract and freelance work. Feel free to reach out to me on my email or social media handles.