Sunil Suresh

Developed and deployed applications utilising state-of-the-art Language Models (LLMs) for both open-source and proprietary frameworks. Proficient in prompt engineering to enhance the performance and specificity of language models. Specialised in content suggestion algorithms leveraging LLMs, enhancing user engagement and personalisation on the platform. Applied LLMs to effectively classify user activities based on messages. Designed an AI-Assessment application which would analyse cases in a conversational manner by asking follow-up questions.

Collaborated with leading organizations for document processing using Natural Language Processing. Designed multi-column page detection and header-footer recognition for legal and lease contracts using heuristics. Developed rules for entity extraction with the help of RegEx based strategies. Developed text classification model for clause classification for Liability Data. Worked on fine tuning BERT model to predict politically biased statements in news articles. Worked on predicting winners for the 95th Academy Awards (Oscars) for 7 leading categories, using sentiment analysis and Q/A modelling. Experienced in developing probabilistic models using pyro utilizing Bayesian Statistics.

Developed a time series analysis of different crops to understand time of pest attacks or other diseases to alert stakeholders. Built machine learning models to predict type of crop and hyper parameter tuning to find best parameters. Worked on data digitization using remote sensing applications like Google Earth Engine and Google Earth Pro. Performed exploratory data analysis of land use land cover using sentinel-1 SAR data.

Minor - Data Science