Pavan Gupta

Experience

• 

  PaperVideo

  Dec 2020 - Jul 2022

  Search Engine Optimization Executive

  Successfully created a chatbot utilizing a Large Language Model (LLM) and LangChain methodology in aneducational technology setting, aiding students in resolving conceptual doubts efficiently. Achieved an 8%reduction in churn rate by enhancing student engagement and satisfaction through the chatbot'spersonalized and effective support.Led the development of an Upsell model using machine learning and deep learning algorithms to executecampaigns on customers, resulting in an 11% increase in retention rate and 4% overall revenue growth.Deployed the model on the Azure platform for scalability and reliability.Conducted detailed analysis on customer behavior and provided actionable business intelligence to higherauthorities for informed decision-making.Assisted the team in automating code and report analysis through pipelines, reducing manual executionefforts by 40%, streamlining workflows, and enhancing productivity.Performed detailed ad hoc analysis using PySpark, SQL, and Python on client requests, enabling swift anddata-driven decision-making processes. Show less

• 

  Msigma Gokulam Pvt. Ltd.

  Jul 2022 - Apr 2023

  AI/ML Project Engineer

  Researching & developing AI/ML models such as DeepOCSORT for object tracking,Vision processing, MMDet for skeleton action tracking, NLP algorithms and creating newAI models using PyTorch to achieve state of art results.Performing ETL operation on big data from AWS S3 to prepare it for AI/ML modeltraining.Communicating with stakeholders, manager, data scientists to understand businessproblems to solve complex problems using advance deep learning framework.Deploying efficiently tested AI model into deployment using CI/CD pipelines bydockerizing the algorithm.Up-to date with latest AI/ML models to make sure solving real world complex problems. Show less

• 

  Hansa Cequity

  May 2023 - now

  Led the team in deploying a Cross-Sell Model using PySpark, combining the XGBoost algorithm and DeepLearning (ANN) techniques. Targeting de-active subscribers, the model identified value-added serviceopportunities, yielding a significant 6.8% revenue increase and 23% customer retention. Executed on a vast datasetof 185 million product purchases, demonstrating scalability and efficiency by following Agile methodology.Engineered,deployed and optimized to an end-to-end recommendation system for DTH clients on Azure cloudusing Azure Kubernetes instances,orchestrating a seamless CI/CD pipeline. Resulted in a 6-8% revenue growth ona dataset of 32 million subscribers.Implemented process automation to streamline data collection, analysis, and reporting processes, leading to a 48%improvement in efficiency and accuracy.Conducted ad hoc analysis on real-time large datasets, enabling timely and informed decision-making for clients.Spearheaded a significant contribution to the LLM (Natural Language to SQL Query) project using the langchainmethod, enabling insightful data analysis by translating natural language queries into SQL queries and extractingactionable insights.Optimized the code latency of stored procedures from 14-16 hours to 3-4 hours, thus improving databaseperformance.Managed the migration of campaign and scoring reports from SQL Server to AWS cloud server end-to-end,including Tableau dashboarding, replacing traditional Excel-based reporting methods.Engineered a recommendation system for a food retail client leveraging LSTM and GRU (NLP) methodologies toforecast the next best item purchase for customers. Achieved an impressive accuracy score of 97.4%, therebyelevating customer satisfaction and boosting sales.Worked in Agile methodology, collaborating closely with cross-functional teams to deliver high-quality solutionsiteratively and adapt to changing requirements effectively. Show less

  • Senior Data scientist | AI/ML

    May 2024 - now

  • Data Scientist | Analyst -Data science and AI

    May 2023 - Jun 2024