Poonam Rath

WORK HISTORY

EDUCATION

ABOUT POONAM RATH

I had my first brush with machine learning during my postdoc at Stanford and fell in love with it. As a wet-lab biologist up until then, I discovered disease mechanism, where my primary mode of being was deep observation, patience and humility. Later on as a data scientist in the startup world, I invented solutions, building on the previously cultivated modalities, absorbing them but transcending them in a sense, with more latitude and chutzpah. In both setups, I have had some common themes: I love questioning, working hard, doing empirical research, and holding space for the fascinating people I have the privilege of sharing my journey with. With a decade of experience working on a variety of use cases in industry, I\'ve built reasonable expertise in weeding out the fluff to get to the essence of a problem space, conceptualizing and then iteratively building solutions that deliver actionable insights (predictions, patterns, recommendations) from unstructured and structured data, as well as complex time series. We’re in an unprecedented era of technological acceleration. It is imperative (and has never been easier) to recognize our flawed fullness of being while also responding to our calling everyday. Scientific training and natural orientation combined, I think my foundational superpower might be my ability to integrate an analytical eye with a heart-centered, intuitive approach to life. I’m intensely interested in being a discerning human, with conscious renewal of sensitivity toward the breadth of the lived human experience in the modern world, and building for the collective good of humanity. My values: Compassion, Humility, Cognitive flexibility, Empathy, Self-leadership, Ongoing authenticity, EquityMy skills (ever-evolving): Critical Thinking, traditional ML, NLP, statistics, linear algebra, deep learning, topological data analysis, Roadmapping and leading teams in cross-functional collaborations, Generative AI, LLMs: prompt engineering, Retrieval Augmented Generation (RAG), LLMOps (serving, routing, load balancing), model training, fine-tuningMy preferred tech stack: Python, Pytorch, Langchain, Langgraph, MLFlow, wandb, SQLEducation: PhD @ Cornell, Postdoc @ Stanford