- M.S., Applied Machine Intelligence | Northeastern University
- B.Tech, Electrical Engineering | Indian Institute of Technology
- Received Bronze medal amongst 200 students by leading a team of 3 members for this 2-week long Datathon using advanced machine learning.
- Achieved F1-score of 0.86533 by incorporating all the steps of the data science solution lifecycle to predict student dropout from courses
- Utilized technologies such as Langchain, LlamaIndex, OpenAI API and the PapersWithCode API to develop an RAG application for researchers.
- Stored indices in a vector database called Pinecone for an LLM Gemini Pro to effectively retrieve information about the latest AI in mental health research papers.
- Developed an end-to-end ML solution to predict whether an existing insurance beneficiary will purchase vehicle insurance.
- Utilized various Machine learning algorithms like logistic regression, decision trees, random forest, XGBoost classifier and LGBM classifier to select the best-performing model and later deploy the solution on Azure cloud.
- Engineered image analyzing tool to provide additional data points about the image by using Instructional tuned LLM.
- Utilized Google Gemini Pro API to conduct prompt engineering and achieve desired results.
- Performed exploratory data analysis about job descriptions from 2020-2023.
- Completed thorough analysis using R to find out the top 10 job titles for an aspirant targeting entry-level full-time jobs.
- Reduced ~$100K in expenses by building interactive Power BI dashboards aiding internal and external stakeholders to make informed decisions.
- Single-handedly developed the firm's entire database by creating an API with a third-party service to schedule batch import at frequent intervals.
- Saved ~$3.5 million in cost by automating IT tasks/processes by building an intelligent automation platform and saved 3+ hours of manual labour for every client with the help of opted services.
- Implemented and maintained APIs and services using Java in collaboration with the UI team which utilized HTML, CSS and JavaScript for the platform to function smoothly.
- Secured a Runner-up position in NeuroHack, an organization-wide hackathon, while leading a 5-member team to create an AI-driven portal for IT ticket categorization and anomaly detection.
- Developed NLP solution and trained BERT model for generating ticket category-2 and incorporated HDBSCAN to generate ticket category-1 from the dataset.
- Developed CNN – Convolutional Neural Network classification model with an accuracy of 96% to differentiate between pizzas.
- Designed an instance segmentation model for biomedical images with an accuracy of 89% (optic disc and fovea detection) and documented a well-designed report and effectively communicated overall results and conclusions of the project.