Saikumarreddy Pochireddygari
Development
New York, United States
Skills
Data Science
About
SAIKUMARREDDY POCHIREDDYGARI's skills align with IT R&D Professionals (Information and Communication Technology). SAIKUMARREDDY also has skills associated with System Developers and Analysts (Information and Communication Technology). SAIKUMARREDDY POCHIREDDYGARI has 6 years of work experience.
View more
Work Experience
Research Assistant
CCDS, SU
November 2022 - Present
- Data Science Duties Developed 25 innovative features for enhancing ML reproducibility using manual coding with MS Excel Spreadsheet Enhanced Geo-data Wiki development platform by streamlining navigation using Python and Docker resulting in good experience Executed large-scale data scraping and analysis of climate-related discussions from 1M Reddit posts and 5M tweets using Python Created a service using LLM, Open AI, Prompt Engineering and Lang chain to identify best Insurance Quote Data Engineer Duties Led thorough data models requirement analysis for the MongoDB to Azure data warehousing migration project Engineered efficient ETL data pipelines with Azure Data Factory, Databricks handling over 2TB of MongoDB data Implemented an incremental data update strategy, ensuring a smooth 2TB data migration with minimal downtime Developed a DAG for unstructured data extraction from application APIs resulting in time spent for data collection efforts by 70%
Machine Learning Engineer
TATA Consultancy Services
June 2018 - August 2022
- Enhanced customer satisfaction to 95.44% through seamless migration of models to AWS, utilizing the AWS tech stack Pioneered a hybrid end-to-end MLOps CI/CD Architecture framework combining AWS IaaS, Databricks Spark Jobs, MLFlow, Snowflake, and Domino Model monitoring to optimize data processing and production model performance in EKS Kubernetes Architected and deployed scalable machine learning infrastructure using AWS CloudFormation, seamlessly integrating with Databricks, and EKS to automate the provisioning and management of resources, leading to a more robust MLOps pipeline Developed and deployed a time series forecasting model to predict customer churn, resulting in a 15% reduction in churn rate. Utilized Prophet for forecasting and implemented an automated pipeline with airflow & Jenkins for model training & deployment Designed custom XGBoost, LightGBM images and developed 3 innovative data pipelines workflows and 15 Airflow routines, reducing infrastructure billable hours by 10% & enabled training in distributed environments with low latency inference Evaluated the feasibility of building models, AI/ML/DL models, model development and A/B tests to predict user adoption for retail client. Explained impact of features on user adoption, informing optimized retention strategies and targeted A/B testing. Crafted Experiments and tested POC prototypes for finance clients, significantly speeding up business deal closures
Education
