Swathy S Aji

Finetuned Large Language Models ( LLM ) like Falcon for tasks such as summarizing tables. Familiar with online machine learning libraries such as River. Developed applications using Streamlit and FastAPI for machine learning models. Performed experiments with Hugging Face ,John Snow Lab nlu and spaCy models across different tasks, which include paraphrasing and NER. Performed experiments in Jupyter Lab, logged my progress with MLflow, and visually displayed the results using various libraries such as Matplotlib. Conducted analysis, data cleaning, feature engineering and employed actual clinical research data to tackle challenging issues within the clinical research domain. Explored, made adjustments to, and utilized datasets that are publicly available for training models. Applied machine learning techniques to effectively utilize modern parallel computing environments. Automated data cleaning tasks to cut down on the time required. Implemented ML solutions to speed up clinical research process. Run machine learning tests and experiments. Documented progress in ML research. Developed scripts to integrate ML models with product. Evaluated models and fine-tune them for business objectives. Retrained the models when there is a model drift. Extended existing ML components in product. Created custom evaluation metrics for specific use cases when required.

Developed an application for training Named Entity Recognition ( NER ) models, encompassing both text-based and document-based models, as well as AutoML Tabular models, and successfully deployed it for production use. Responsible for server that offered all AI-based services to the RPA studio. Experience in h2o, spaCy , tpot ,NLTK and hugging face models like LayoutLM. Understood Business Objectives and Developed the ML Models as per need. Performed Machine Learning Model Tests and Experiments. Experience in working with Linux OS ,Windows ,virtual machines and Google colab. Experience with creating web APIs in Flask-RESTful. Experience with ML libraries like Pandas, Numpy, Seaborn, Opencv, Scikit-learn and Matplotlib. Experience in professional software development practices and tools for source control and project management like Git and JIRA. Experience in Sql Alchemy , Sqlite , regular expressions ,and Azure Form Recognizer. Developed model that can translate sentences from European languages to English using word vectors. Created dataset (including data labeling ) and trained key-value pair extractor model for invoices. Labeled data, supervised dataset splitting, and developed NER model for shipping-related emails.

Used different ML algorithms to solve problems. Experience in Python, Sqlite and Tkinter. Experience in image processing using OpenCV.

IOT based vehicle parking manager , makes finding parking slots more efficient ,less time consuming for the employees . Smart parking includes the use of low cost sensors ,realtime data and android application that allow users to monitor available and unavailable parking spots.