Research Software Engineer
Current• Research, develop and publish on different projects taken up by the Software Directorate at NCSA.• Developed a customizable and scalable open-source data management platform (clowderframework.org) hosting GPT models for sentence and document classification in a Kubernetes cluster. Developing intelligent document parsing based on LLMs to automatically assess quality of medical publications and give feedback to authors and reviewers as to whether the paper meets the review criteria.• Implementing agents in LangChain for writing and executing RNA Sequencing analysis pipeline. Uses Github webhooks for human interaction, Plan and Execute agent workflow with OpenAI and Azure GPT4 models, LangGraph for async agent action control, and Ray to run distributed Python workloads.• Developing a platform for distributed ML training and inference on HPC clusters hosted by NCSA.• Developing RAG agents to analyze email dataset and explore the reasoning behind changes to the business model for SHEILD Illinois Initiative to make the innovative saliva-based COVID testing available to the public.• Developing a platform for distributed ML training and inference on high performance computing clusters hosted at NCSA.• Backend platform design and development on Rokwire (rokwire.org) - an interdisciplinary project at University of Illinois for building an ecosystem for smart, healthy, and connected communities.• React-Redux web application for an open-source data management platform for National Institutes of Health Randomized Clinical Trial publications. Topic classification using BERT and GPT based models.• Automation of data processing pipeline to support submission of open-source weather data (Ameriflux ameriflux.lbl.gov) from crop sites.• Member of ML Special Interest Group - created study materials for RSE's to learn ML. - led Round Table discussions & Hands-on sessions on ML & RAG.• Conferences attended : - NeurIPS 2022 - AGU 2022 - US-RSE 2023 - US-RSE 2024