Sudharsan Ravichandiran

Leading the Gen AI Projects. Created a FAQ bot by fine-tuning SOTA open-source LLM with RAG. Built a robust LLM framework that serves as a playground for fine-tuning various LLMs across tasks enabling rapid iteration and optimization. Built an intelligent software testing platform that performs unit testing, automation and manual testing using LLM. Built an interview bot using LLM that can conduct prescreening interviews and evaluate candidates. Worked on preventing LLM abuse using Guardrails.

Led the generative AI initiatives. Contributed expertise in prompt engineering and prompt tuning for state-of-the-art generative AI models. Developed a document answering model using LLM with langchain. Built a smart invoice information detection on the medical records using Layot LM.

Led a team of data scientists and all data science initiatives. Responsible for all data activities right from data collection to model deployment and monitoring. Drive the data science roadmap, and collaborate with various cross-functional teams. Developed a grammar error correction model for correcting grammar errors in the documents. Built a recommendation system using a vector space model for assigning relevant reviewers to review the documents. Worked on query expansion using BERT for enhancing the search results. Worked on keyword trend analysis to identify the pattern in the searches made by the user. Developed a document question-answering model using BERT for answering questions from a huge collection of documents. Developed a next sentence prediction model using GPT-3. Devised a smart document tagging model for categorizing and organizing documents.

Authored a best selling book titled Deep Reinforcement Learning with Python which covers from the classic reinforcement learning algorithms to advanced deep reinforcement learning algorithms. Worked on independent research which includes GAN for imitation learning with expert trajectories. Authored a Deep Learning Pocket Reference Guide with TheDataGuy series which includes frequently asked deep learning questions.

Led data science initiatives and coordinated with several cross-functional teams. Developed a deep learning model that understands the hiring behavior of the recruiter and stack ranks the resumes accordingly. Built a text summarization model for summarizing the resume using encoder-decoder architecture. Prototyped a personalized recommendation system using deep reinforcement learning with trust region policy optimization. Handled relatively huge social datasets and built a statistical model for deriving social intelligence. Built several analytics dashboards for deriving latent insights which include competitive intelligence, ATS intelligence, and more.