AI Vacation Manager with Llama-2 with MLOps and Prompt Eng

In this project, I deployed Llama-2 on AWS SageMaker to create an innovative AI Vacation Manager, employing deep learning containers for seamless integration.

I utilized advanced prompt engineering techniques, including the Chain of Thoughts and Tree of Thoughts, to guide the model through complex reasoning, ensuring responses were tailored and practical for vacation planning.

By optimizing key parameters like Temperature, top_k, and top_p, I fine-tuned the model's output, striking a balance between creativity and relevance.

This approach not only enhanced the user experience by providing diverse, customized vacation planning suggestions but also showcased the potential of AI to simplify and improve decision-making processes in everyday tasks.

Huninchen Chat Bot: Leveraging LangChain on Azure

In the Huninchen Chat Bot project, I engineered an intelligent question-and-answer bot using Azure, Streamlit, and OpenAI technologies, focusing on interactions with a fictional entity named Huninchen.

This bot utilized Azure Cognitive Search for sophisticated querying capabilities, Azure Container Registry and App Services for robust deployment, and Azure Blob Storage for efficient data management.

By leveraging the LangChain framework, I was able to create a conversational agent that not only understood and responded to queries about Huninchen with high accuracy but also demonstrated the integration of complex cloud services and AI models to deliver a seamless user experience.

This project exemplifies my ability to harness cutting-edge technology to create innovative solutions that enhance digital interactions.

Game deployment on EKS with Fargate compute

I took on the challenge of deploying the popular 2048 game on AWS from the ground up, utilizing the Elastic Kubernetes Service (EKS) with Fargate for the computing aspect.

This endeavor required a deep dive into Kubernetes management and AWS services, where I skillfully used tools such as kubectl, eksctl, and the AWS CLI to orchestrate the deployment process.

A significant part of the project involved configuring a load balancer to ensure the game was accessible over the internet, showcasing my ability to implement scalable and accessible web applications.

This project highlights my expertise in cloud computing, container orchestration, and my commitment to leveraging advanced technologies to deliver engaging and scalable web solutions.

Global Health Information System

In this innovative project, I developed a Global Health Information System, a specialized search engine designed to aggregate and analyze global health news.

This system was meticulously engineered to scrape healthcare information from various websites, with a keen focus on identifying disease outbreaks and providing up-to-date information on medical products and research.

By utilizing the MPNET model for text embeddings and document similarity, along with pretrained Q&A models for entity extraction, I ensured the system could efficiently gather and present relevant health information in an organized manner.

This project underscores my commitment to applying AI and machine learning technologies to address real-world challenges, specifically in the crucial domain of global health.

Spotify Data Analytics: Automated ETL Pipeline and Visualization on AWS

In the ETL Pipeline project for Spotify songs data, I automated data warehousing on AWS to streamline data handling for an IAM user.

By leveraging S3 for storage, AWS Glue for visual ETL processes, and Crawlers to create tables, I enhanced data organization and accessibility.

Utilizing Athena for data querying and QuickSight for visualization, I provided insightful analysis and interactive data representations, showcasing my ability to harness AWS services for efficient data management and analysis.

ERP tool for Hadoop cluster setup

In the project "ERP Tool for Hadoop Cluster Setup," I developed a user-friendly interface that allows users to effortlessly configure and deploy Hadoop clusters on Docker containers, leveraging Ansible for automation.

This tool not only streamlined the setup of HDFS and Hive services but also included a feature for performing Twitter Sentiment Analysis, enhancing the analytical capabilities of each Hadoop cluster created.

This solution significantly simplified the process of big data management and analysis for users, demonstrating an innovative approach to deploying and managing big data technologies.

Online Songs Player

In the "Online Songs Player" project, I developed a fully functional online music player as part of my J2EE training, utilizing a combination of Servlets and JSP for the backend and HTML/CSS/JavaScript along with Bootstrap for the frontend design.

The project featured a robust backend that managed song data storage in an SQL database, with efficient data retrieval facilitated by Hibernate.

This approach provided a seamless user experience for accessing and playing songs, showcasing my ability to integrate various web technologies to create a comprehensive web application.

Recommender System

In this project, I developed an advanced song recommendation system that employs a hybrid method combining both user-based and item-based collaborative filtering techniques.

This system effectively personalizes song suggestions by analyzing the listening habits and preferences from a comprehensive user-songs dataset.

To address the challenge of recommending newly added songs, I integrated a mechanism that suggests popular tracks to users, ensuring that even the latest additions receive visibility.

This approach demonstrates my capability to innovate within the realm of recommendation systems, providing users with a highly tailored listening experience.