Shivendra Singh

Project : Sales Representative Chatbot

• Developed chatbot to assist medical sales representative.
• Implemented NLP models to handle complex queries and provide accurate, role specific responses.
• Integrated multiple expert agents (Legal Counsel, Medical Expert, Marketing Agent, Graphic Design Agent).
• Created a feedback mechanism to improve chatbot responses using user feedback and Pinecone for vector storage.
• Enabled file uploads for real-time data updates and insights.
• Designed a user-friendly React-based web interface for seamless interaction.
• Technologies: React, Flask, Pinecone, SentenceTransformer, Axios, CSS, Tokenization, Embeddings, Similarity Search, Data Preprocessing, Multimodal Data Handling, RESTful APIs, Feedback Loop

Project : V2X

• Worked for the client Suzuki on the V2X project.
• Designed and developed a ZMQ-based pub-sub system with Boost C++ libraries and created use cases like Road Hazard Warning, Assisted Overtaking Maneuver, and a communication system using Vanetza.
• Automated geo-simulation with launch scripts. Integrated vehicle-to-everything tech into autopilot, enhancing ACC, ALC, acceleration, braking, and lane following.
• Implementation of Level 2 autonomy with V2X. Tech Used: Python, C++, Linux, ZMQ, Vanetza, Machine Learning, Smart Pointers, Shell Scripting, Git.

Project : DApp (Road Monitoring)

• Spearheaded the implementation of image processing capabilities in the app for obstacle and signage identification, ensuring data authenticity.
• Edge computing for real-time information processing and integration of cloud upload functionality for storing footage, enabling more intensive hazard and obstacle classification. Tech Used : Python, Deep Learning, Tensorflow, Bash Scripting, Git, Linux, Spring Boot

• Designed and developed a SMART Neonatal Incubator and non-contact measurement of infant vital parameters and Bilirubin using Image Processing, Machine Learning, Python
• Designed and developed an application for tooth denture analysis using Python, GUI toolkit and pyinstaller. Tech Used: Python, Linux, Machine Learning, Image Processing, GUI toolkit, Algorithm design, Rasp OS.

• Develop an application to identify the walls of common carotid arteries for early detection of cardiovascular diseases.
• Development phase included data collection, data cleaning, feature extraction, testing of ML models and testing of device in real time. Tech Used: Python, Machine Learning.