H Vilas

• JSW Safety Prediction Analytics: Working on React.js frontend using react-grid-layout for dynamic graph layouts.
• Developed backend using Django and integrated with LLaMA to build accident and safety prediction models for JSW clients.

• Fullstack Developer: Built REST APIs using Django and Django REST Framework, improving query performance by 30%. Designed responsive UIs with React, integrated APIs via Axios, optimized frontend with lazy loading and caching.
• DevOps: Deployed with AWS EC2 using Terraform and Docker. Implemented CI/CD with Jenkins and SonarQube for code quality.
• Community Health Worker App: Developed app for CHWs to register households, auto-generate e-vaccination forms, and provide condition-based medication advice. Collaborated with Swiss client and received appreciation for backend API work using JavaScript.

• Backend Development and Research: Worked on a Kafka-based data streaming system for client Tenxlabs, enabling real-time processing and analysis of sensor data with backend techstack Fastapi

• ML and NLP Projects: Built stock price prediction models using scikit-learn and TensorFlow. Worked with LangChain and fine-tuned LLMs with LoRA/QLORA. Integrated Pinecone for vector search in chatbot applications.

Designed and developed a personal portfolio website using React, TailwindCSS, and deployed it on Vercel. The portfolio showcases my projects, educational background, achievements, and skills. It is accessible at https://portfolio-delta-navy-47.vercel.app/, providing a comprehensive overview of my professional and academic journey.

(PyTorch, TensorFlow) Engineered a conversational AI model that predicts contextually appropriate next words using deep neural networks. Optimized with backpropagation, ReLU and hyperbolic activations, negative log-likelihood loss, Kaiming initialization, and batch normalization to reduce gradient issues and enhance accuracy.

(Full-Stack) Built a robust full-stack application using Django and React.js. Implemented JWT authentication and dynamic REST APIs.

(RAG, Langchain, Langserve): Developed a tool that recommends restaurants based on city input using natural language processing for accurate suggestions.

(Front-End) Built a responsive music player app in React.js using hooks like useState and useEffect, with intuitive controls, modern design, and audio playback features.

(Machine Learning) Created a regression model using scikit-learn to predict house prices based on area, including feature engineering for accuracy.

Credentials: STB99412

under lecturer Matt Constable

Volume 13, Issue 4, April 2024. DOI: 10.17148/IJARCCE.2024.1341203