• Contributing to AI product development through prompt engineering, langchain, and tool building, enhancing user insights into their data leveraging Teradata along with results evaluations using metrics like Ragas, Vectara, and employing LLMS• Exploring open-source alternatives such as LLama and implementing faster execution using Streaming LLMs. Additionally, quantizing models using GPTQ and GGUF to enhance model performance, and fine-tuning using Lora to improve responses

• Responsible for developing SIT workloads for In-Database (InDb) Analytical Functions, encompassing a wide range of tasks, including Time series analysis, Supervised learning, Unsupervised learning, NLP, and various data preparation functions• Generating value by conducting comprehensive benchmarking of Teradata analytical algorithms against Python and R open-source implementations, identifying their strengths and weaknesses in terms of performance and accuracy• Development of POCs for customer use cases, utilizing Teradata's diverse analytics products to showcase their effectiveness in real-world scenarios. This includes analyzing large amounts of data to provide insights and actionable items to drive business decisions

Graduate Teaching Assistant for Natural Language Processing course.

• Ensuring safety of students at their residence halls by creating a safe environment with rules.• Coordinate with the RSO Supervisors, NUPD, Resident or an RA in case of emergency.• Help students in their difficult times such as lockouts, directions, lost property, etc.

• Created end-to-end ML use cases for Telco Churn, Heart Failure, Credit Card Bust out data to demonstrate partner integration capabilities of Teradata Vantage for tools such as Dataiku, H2O.ai, Sagemaker, and Vantage Analytics Library• Demoed built uses cases to internal teams like marketing, customer success, engineering, and customers like AT&T• Devised a BYOM (Bring Your Own Model) plugin for Dataiku to export the ML models built in Dataiku as PMML to Teradata Vantage tables and score models• Cleansed and analyzed Teradata’s last six months telemetry data (~9M rows) by parsing JSON columns, created dashboards using Tableau and Power BI, identified focus areas to increase number of customers for ML products

• Designed machine learning models using K means and Hierarchical Agglomerative clustering algorithms for drug consumption data to evaluate the risk with over 80% accuracies• Responsible for communicating with clients to understand their requirements and worked with team to solve problems by meeting deadlines• Collaborated with team as a part of training and learned about analyzing and visualizing data utilizing R and Python by using ggplot2, seaborn libraries

• Lead a team of 5 to develop machine learning models SVM, Logistic regression, ANN for the credit card dataset to detect fraud transactions with test accuracy 97% (Logistic regression) by producing balanced dataset using SMOTE• Interacted and worked with HPE professional to understand HDFS, Cloudera, Map Reduce by working on Linux (CentOS)

• Enhanced knowledge about its application architecture• Developed Throttle Auto Solutions webpage with the team and learned about its e-commerce framework