Shivam Bajpai

• Led team of data scientists to successfully complete complex projects VERTEX which has LLM powered search capabilities (text2SQL) and Tailored Risk Analysis models providing accurate risk predictions to network data.
• Worked on creating Knowledge Graph from relational DBs and LLMs
• Successfully issue of patent -INTERACTIVE REAL-TIME VIDEO SEARCH BASED ON KNOWLEDGE GRAPH
• Developed on MVP for troubleshooting user query by uploading product image real time on chatbot
• Currently pursuing MTech Specialisation from IISc online

• Developed ML/DL deployment platforms and MLOps solutions, streamlining the model deployment process and enhancing operational efficiency.
• Engineered an automated deployment workflow utilizing customized Python applications to package and deploy machine learning models effectively.
• Spearheaded the development of tools for model monitoring and distributed training, ensuring optimal performance and scalability across the organization.
• Collaborated closely with computer vision (CV) teams to implement theft detection deep learning models, contributing to enhanced security measures.
• Designed and implemented dashboards on Tableau to visualize performance metrics and deployment timelines for machine learning models, facilitating data-driven decision-making processes.

• Led the development and implementation of a gesture-based solution for favorite pose classification using Deep Learning models in PyTorch, deployed in luxury Mercedes Benz cars, processing over 1 crore images.
• Created interactive dashboards in Tableau to visualize charging station data, providing actionable insights for stakeholders.
• Designed and executed a solution to predict driver and passenger body joints (keypoints) from interior camera grayscale images, enabling gesture-operated functionalities.
• Optimized model hyperparameters and reduced training time using PyTorch Lightning framework.
• Developed and maintained routing services in Java and Python for electric vehicles, dynamically integrating charging station locations based on battery state of charge for efficient route planning.
• Implemented machine learning models to forecast energy consumption and estimate travel distance based on the state of charge, seamlessly integrated with routing services.
• Engineered Python algorithms to interface with route content provider systems across different markets, delivering region-specific routes for vehicles based on their location.
• Orchestrated the migration of services to Azure Kubernetes, establishing robust deployment pipelines and Continuous Integration/Continuous Deployment (CI/CD) practices.
• Spearheaded a Proof of Concept (POC) for a smart voice assistant integrating multiple voice platforms (Alexa, Google Home) for enhanced user experience and accessibility.

Hotel Reshop Voice enabled chatbot (American Express)

Description : Developed a chatbot using open source platform , RASA core . The customized actions are written using python which in turn call Microservice to get details.

Hotel Reshop Artificial Intelligence Data parser & Micro Service (American Express)

Description: Developed an AI solution using RASA NLU and Dialog Flow (api.ai), that parses raw data which contains hotel amenities details from itinerary XML , which is further used to provide recommendation to the customer . Developed a production level data parser using RASA NLU which works with an accuracy of 97%

Technology Used:

RASA NLU, Openshift, Docker , Dialog flow , Java 8 , Spring boot ,RASA CORE,

Global Document Delivery- Trip details, Itinerary & E-ticket micro service Application :

This project was introduced in Global Business Travel to replace the existing system where we receive the itinerary & e-ticket xml.

Environments Used: Java 8, spring boot Log4j, Apache tomcat, Microservices, Docker, Maven

Digital Travel Record Application :

American Express Digital Travel Record (DTR) system is a back office application used by Travel counsellor to process a PNR from Global Distribution System (GDS).

Technology Used:

Core Java, SOAP and REST web service, SQL server 2008,