●Developed and deployed AI prototypes within the Healthlytics.ai division, leveraging LLMs, TTS, and computer vision models to create real-world solutions for healthcare automation.
● Managed deployment of AI models and services to cloud environments using Google Cloud Platform (GCP) and Microsoft Azure, including configuring virtual machines and managing compute resources.
● Designed and implemented a full-stack web application to interface with AI models, building responsive front-end dashboards and scalable backend services that handled over 1,000 API requests daily, enabling real-time model inference and reducing data processing latency by 40%.

● Collaborated with Happy Nutrition’s Health Academy to enhance their online learning platform built on WordPress, improving usability, SEO, and overall user experience.
●Redesigned and optimized course layouts for better navigation and learner engagement, integrating slides, modules, and multimedia content.
●Ensured payment gateway functionality and seamless enrollment processes, making the platform more accessible to users.
●Applied SEO best practices (keyword optimization, meta descriptions, performance improvements) to increase visibility and search rankings.
●Improved site responsiveness and accessibility, ensuring a consistent user experience across desktop, tablet, and mobile devices.
●Implemented marketing and promotional features to support course outreach, including landing pages, lead capture forms, and social media integration.
●Collaborated with mentors and stakeholders through regular check-ins, gathering feedback and iterating on design and functionality.
●Completed and launched 1–2 fully functional online courses, including structured content and interactive learning materials.

●Researched and analyzed Hugo’s static site framework, exploring existing open-source templates, features, and limitations to determine suitability for different industries.
●Conducted a needs analysis to identify requirements for target industries such as hospital departments, DJ/photography services, industrial companies, and universities.
●Designed and developed industry-specific Hugo templates, customizing layouts, navigation, and content structures for professional use cases.
●Deployed website prototypes using Cloudflare and GitHub, ensuring reliable hosting, fast performance, and version-controlled development workflows.
●Authored technical documentation and training guides with clear step-by-step instructions for non-technical users to independently update and maintain websites.
●Contributed to the redesign of the company’s main website, improving responsiveness, usability, and visual branding.
●Participated in an iterative feedback process with mentors and stakeholders, refining templates and functionality based on real-world needs.
●Gained hands-on experience with website architecture, template customization, deployment pipelines, and static site optimization.

● Developed an AI agent using LangChain to automate searches, compare results, and extract relevant data, streamlining information retrieval and reducing manual effort significantly.
● Built a vector store database using Astra DB to optimize client data access, improving query performance and retrieval speed by over 50%.
● Enhanced data processing workflows, leading to faster and more accurate extraction of client insights from large datasets, improving efficiency across key operations.

● Engineered and maintained custom scripts and managed website operations, achieving a 30% improvement in page load times and a 20% increase in user engagement.
● Designed and deployed an intranet system and optimized databases, reducing data retrieval times by 40% and enhancing internal communication efficiency.
● Automated workflows using cloud flow automation, reducing manual processing time by 50%, leveraging technologies such as JavaScript, HTML, CSS, Microsoft FX / Flow, SharePoint and Python.

● Developed a machine learning algorithm for one-shot classification of handwritten character images using Modified Hausdorff Distance, achieving reliable similarity measures with minimal training data.
● Implemented image processing and distance calculation functions to convert images into arrays of inked pixel coordinates and compute the MHD between them.
● Executed 20 test cases to find the prediction error rate, calculating the average error rate across multiple runs to evaluate the system's accuracy and effectiveness in image recognition tasks.

● Built an end-to-end offline image classification app using a custom CNN in PyTorch, achieving 87% test accuracy across 10 Korean dishes with sub-200ms local inference time.
● Engineered a deep CNN architecture with residual-style blocks, adaptive pooling, and dropout, trained using the AdamW optimizer and dynamic learning rate scheduling for robust performance.
● Applied six custom data augmentation techniques (e.g., mosaic, Gaussian noise, histogram equalization) to expand a dataset of 1,200+ labelled food images, improving generalization.
● Designed a Tkinter-based desktop UI for image upload, classification display, and food description.

● Developed a secure login system handling 1,000 daily authentications with encrypted credential storage.
● Created an interactive member dashboard with asynchronous data fetching for real-time updates,
including membership status, attendance, and payment history, achieving a 95% user engagement rate.
● Designed an admin interface for managing 1,000+ user metrics, utilizing complex SQL queries to streamline analytics for payment, delinquency, and revenue.
●Employed Agile methodology throughout the development cycle for iterative improvement and rapid feedback integration

● Conducted in-depth analysis of student feedback from a structured survey, emphasizing the successful integration of AI-based educational tools like ChatGPT in academia, resulting in 75% higher response rate.
● Utilized Python, particularly Pandas, for efficient data management, cleaning, and analysis, ensuring data quality and reliability.
● Applied advanced statistical techniques and created data visualizations using Matplotlib and Jupyter to extract valuable insights and support data-driven decision-making

● Developed advanced Java audio application, integrating robust library management, download and playlist features, and an efficient File I/O system to maintain a large, updated content library.
● Improved error handling by introducing custom exceptions, resulting in a 40% boost in managing exceptional situations and achieving 99% test coverage.
● Enhanced user experience with a 25% reduction in navigation time through an intuitive search feature for easy content discovery by title, artist, or genre.