Data Scientist Profile: Vishnu Vandith Guntupalli

• Vishnu Vandith Guntupalli is pursuing Data Scientist opportunities within the IT industry.
• He is passionate about using natural language processing techniques for innovation.
• He is open to global opportunities.
• Contact: [email protected], +91 7207776177, Hyderabad, Telangana, India.
• Skills include Machine Learning, Deep Learning, Unsupervised Learning, and Time Series analysis.
• Programming languages: Node.js, Python, JavaScript, and others.
• Frameworks and APIs: TensorFlow, Keras, Sklearn.
• Core competencies: Analytics Roadmap, Data Analytics & Governance, and more.

Profile Summary

• He has over 3 years of experience as a Data Scientist, focusing on Large Language Models (LLM).
• Expertise spans data collection, analysis, model development, and deployment, especially with LLMs.
• Skilled in creating and deploying machine learning models, including LLMs, for natural language understanding.
• Experienced in extracting insights from complex datasets using LLMs for text analysis.
• Proficient in advanced analytical techniques, using LLMs for predictive analysis and customer profiling.
• He applied Retrieval-Augmented Generation (RAG) to enhance Large Language Models.
• Leveraged MLOps principles for streamlined model deployment using CI/CD pipelines.
• Integrated monitoring and logging to track model performance.
• He implemented automated model retraining and versioning to maintain LLM accuracy.
• Utilized cloud infrastructure for scalable model deployment.
• He collaborated to integrate MLOps practices into the data science workflow.
• Adept at using NLP techniques powered by LLMs for textual data insights.
• He is proficient in Python and libraries like Pandas, NumPy, and Matplotlib for text data processing with LLMs.
• He is known for developing business solutions using LLMs to improve revenue and cost-effectiveness.
• Capable of optimizing data collection, especially for text data, supporting analytical systems with LLMs.
• He is skilled in understanding client needs and designing solutions using LLMs for text-based tasks.
• Experienced in designing and deploying data-driven applications with LLMs for natural language understanding.
• He built prototype applications using Large Language Models (LLM) and advanced vector search techniques.
• Fine-tuned machine learning models on custom data.
• Deployed developed solutions into production.

Work Experience

• Data Scientist at Transloom Inc (Hyderabad, Telangana, 09/2023 - Present):
• Developed a FastAPI-based translation service using pre-trained NLP models.
• Implemented a scalable architecture using asynchronous programming and preloading NLP models.
• Employed Jenkins-X for CI/CD automation, streamlining build, test, and deployment.
• Utilized pre-trained NLP models like MarianMT and Facebook's M2M100.
• Integrated Swagger UI for API documentation.
• Implemented model monitoring and logging on AWS using CloudWatch.
• Automated model retraining and versioning using AWS S3 and Jenkins-X.
• Deployed scalable infrastructure on AWS Elastic Kubernetes Service (EKS) for model serving.
• Addressed complex issues related to model loading and deployment.
• Data Scientist at Openstream.ai (Hyderabad, Telangana, 06/2021 - 08/2023):
• Established a pipeline for converting natural language expressions into elastic queries.
• Designed services for language detection, translation, and elastic data loading.
• Enhanced document extraction for user-uploaded documents, improving Chatbot query management.
• Used Beautiful Soup and Selenium to scrape news data from The Times of India.
• Developed an object detection model for identifying table types in PDF files.
• Integrated bi and cross encoder models for semantic search functionality.
• Adapted the API to use Langchain tools through the Langchain Agent.
• Fine-tuned LLM models on custom data using QLora and Lora.
• Constructed an end-to-end pipeline for document ingestion and querying using Langchain and LLM.

Key Projects

• EvaAir:
• Built applications using LLMs and vector search for data extraction from PDFs.
• Fine-tuned machine learning models on custom datasets.
• Expertise in Python ecosystem, using FastAPI, PyTorch, NumPy, pandas, and sentence-transformers.
• Proficiency in deploying solutions to production using Docker, git, Linux, and bash.
• Developed a tool for extracting information (names, addresses, amounts) from PDF documents.
• Leveraged LLMs for text processing, and implemented a user-friendly interface for PDF data extraction.
• Bot Framework:
• Implemented capabilities in a multimodal framework using Sentence Transformers for contextual understanding.
• Leveraged Sentence Transformers for empathetic bot responses.
• Created chatbots capable of answering queries, providing recommendations, and processing transactions.
• Extended the framework with image recognition using LLMs.
• Integrated voice-to-text functionality with LLMs.
• EvaMHA:
• Developed an application for real-time facial expression detection.
• Developed a mental health application using Power BI for real-time data visualization.
• Designed custom data connectors to integrate data from wearables and surveys into Power BI.
• Implemented DAX calculations to analyze changes in emotional states.
• Collaborated with mental health professionals to tailor Power BI dashboards, using LLMs for content generation.
• Facilitated the diagnosis and monitoring of mental health conditions through Power BI and facial expression analysis.
• Protopy:
• Developed a Prolog-to-Python parser tool to simplify development.
• Designed customizable user interface used in NLP and AI applications.
• KGQA (Knowledge Graph Question Answering):
• Implemented a Q&A system using Neo4j for accurate answers from a knowledge graph.
• Developed a knowledge graph representing concepts, entities, and relationships.
• Utilized Neo4j and Cypher for querying graph data.
• Integrated NLP techniques for understanding user questions.
• Used named entity recognition, part-of-speech tagging, and semantic parsing.
• Plan-based Agent:
• Developed a user interface (UI) for a plant-based agent using Angular.
• Created an intuitive UI design to enhance user experience.
• Implemented user authentication and authorization features.
• Skills as a Intern:
• React Native Developer (MegaHoot Technologies, Inc): Developed a hybrid video calling app with blockchain security.
• Intern (Produens Labs): Developed an Atlassian plugin using Node.js and React for JIRA and Confluence integration.
• Intern (Indian Servers): Implemented a decryption tool for ransomware attacks and conducted website threat analysis.

Community Experience

• Google Developer Students Clubs VJIT | Tech Lead
• Code Chef VJIT | Competitive Programming Lead
• IUCEE Annual Student Forum: (Student Co-Chair/Web Developer)

Sessions

• National Level Hackathon Jury Member
• ML Bootcamp
• Angular Bootcamp
• Web Application Exploitation and Security
• MongoDb with Ease (upcoming)

Research

• Identity and Access Management using Artificial Intelligence
• Multi-Intent Classification Using Dependency Parsing and Named Entity Recognition
• Design And Development Of Water Quality Monitoring System In IOT:

Patent(s)

• NLP interpreter with display device