Where Ideas Meet Impact: Discover My Professional Path

Spark AI is a platform designed to empower content creators, especially influencers and niche creators, by streamlining the content creation process. With the explosive growth of the influencer marketing industry and the increasing adoption of AI, there is a need for tools that cater to the specific challenges faced by creators.Through interviews and surveys, we identified key pain points for creators, including the struggle to grow engagement, the overwhelming number of AI tools disrupting workflows, and the need for a unified platform that seamlessly integrates text and image generation, referencing, and content posting.Spark AI aims to address these challenges by providing a one-stop solution that leverages the power of Gemini AI. With a single prompt, Spark AI generates niche content tailored to the creator's needs, along with relevant references to build trust with the audience, accompanying images, and the ability to directly post the content to a medium blog.

Our product, developed by a dedicated team of 7, leverages advanced U-Net ML based architecture to transform MRI scans by reducing redundant motion artifacts found in 30% of cases. This innovation not only cuts down the cost and time associated with acquiring different MRI modalities but also enhances diagnostic accuracy with a remarkably low mean squared error of 0.00566. Furthermore, our models have been successfully deployed to a user-friendly website, streamlining access for medical professionals.
This project was recognized internationally, winning the CSI International Applications’21 Award.

Sequio redefines test preparation, merging AI with a gamified learning experience tailored to individual styles, enhancing efficiency. Our development, informed by significant user research (50 interviews and 400 surveys), led to strategic, need-based agile sprints, introducing impactful, user-focused enhancements. Our AI-driven approach to question selection and personalized learning pathways has doubled learning speed and retention, earning us accolades like the Silicon Valley Innovation Challenge and interest from industry partners.

https://blogs.sjsu.edu/newsroom/2023/silicon-valley-innovation-challenge-kicks-off-its-20th-year/

This project is a visual tech-education tool in its traction phase, aimed at demystifying technology for all ages. This mockup, created to gauge customer interest, has seen positive traction through Facebook ads and feedback from our demo website surveys, signaling strong market potential and user engagement. "Help Me Out" is a concept created through extensive user surveys, market research, and insights gained from advertising analysis. The goal is to develop VideoGPT style platform customized for the older generation, designed to assist them in using technology for their everyday tasks.

For a preview of the website mockup, click below.

This project uses machine learning to predict liver disease risk in Patients, focusing on factors like alcohol and drug consumption. Utilizing data from the UC Irvine Repository, We have applied mutiple ML models like Regression Tree, Classification Tree, KNN, Gradient Boosting, Random Forest, and selected classification tree for its low misclassification rate and a high AUC, aiding in accurate detection and management of liver issues.

The aim of the project to compare different models, tune them and pick one for accurate predictions.
 
 

Our "Covid-19 Prediction" project aims to analyze global impacts, highlighting continent-specific case contributions, death rates per million, and case numbers per million. It explores variable correlations, benefiting societal understanding and aiding further research. Data, sourced from Worldometer via Kaggle, covers February 2020 to March 2022. Findings reveal strong correlations between deaths and testing, with Europe noted for high death rates. This contributes to understanding pandemic dynamics, with future work potentially delving deeper into variable influences and regional analyses.

"Evanto" is an innovative Android app designed for event planners, integrating AI to predict event success factors like weather and traffic. It offers secure user management and data handling through Firebase API, with Razor's payment gateway for smooth transactions. Enhance your event planning with Evanto's predictive analytics and streamlined digital features.

In our paper, "Multi Modality Medical Image Translation for Dicom Brain Images," my team and I introduce a novel approach to medical imaging, specifically aimed at enhancing the efficiency of MRI scans. By employing a deep learning model based on the U-Net architecture, we have developed a method to quickly and accurately convert between two common types of MRI images: T1-Weighted and T2-Weighted. This conversion is vital because each type of image provides unique diagnostic information, but obtaining both is typically time-consuming and expensive.

Our approach significantly reduces the time needed for these scans, from several minutes to just seconds, thereby improving the patient experience and reducing costs. We've tested our model extensively, and the results show that our method outperforms existing solutions in terms of accuracy and efficiency.

The implications of our work are vast. Not only does it have the potential to make MRI scans more accessible and less burdensome for patients, but it also opens the door to applying similar techniques to other types of medical images and different MRI modalities. We believe our research marks a significant step forward in the use of deep learning within medical imaging.

In the ever-changing world of Business, making a successful product design requires quickly trying out new ideas and listening to what users think. With the integration of Artificial Intelligence (AI), the process of rapidly generating content from existing data becomes easier, accelerating the creation of innovative product concepts. To use this cutting-edge technology to build novel products a framework is essential. The paper aims to propose a novel framework that caters to a customer centric integration of AI in Product Designing.  The paper goes beyond theoretical explanation, implementing the framework to iteratively design 3 diverse products. The implementation also highlights the potential opportunity and challenges of using AI enabled framework. Finally, the paper concludes with the critical insights obtained with the implementation of the framework.