Shubham S. Email & Phone Number

- Worked on developing a cutting-edge thematic analysis engine powered by Large Language Models (LLMs) such as GPT, BERT, and Instruct models. Utilized language model finetuning, unsupervised clustering, and sentiment analysis techniques to enable nuanced data interpretation across a variety of domains, improving insights for key stakeholders. Developed an open playground for qualitative studies, including In-Depth Interviews (IDIs) and focus groups, enabling flexible exploration of qualitative data. This solution played a central role in delivering significant value to the customer by enhancing their ability to extract actionable insights from complex data.- Orchestrated the migration of compute-intensive, queue-based services to a microservices architecture, leveraging AWS Lambda for efficient serverless function execution and AWS Fargate for robust container orchestration, resulting in improved scalability and reduced operational costs. Additionally, played a key role in setting up infrastructure with enterprise-grade security, implementing AWS best practices such as IAM role-based access controls, VPC configurations, and encrypted data storage. This ensured the environment was both highly secure and compliant with industry standards while maintaining optimal performance and flexibility.- Developed data processing pipelines for media content that included tasks such as image upscaling, transcription, and speech diarization. Implemented data sanitation measures, including voice and facial activity detection. - As a full‑stack developer, ported an existing web application to run natively on Android devices, enhancing its accessibility and user base. Also engineered offline functionality into the app to ensure usability in areas with low or no network availability.

- Engineered a User Streak feature to monitor user engagement metrics over time, helping users to better engage with the platform, resulting in more questions solved per month- Implemented a strategy to decrease user effort per topic by eliminating redundant questions, leveraging historical analytics data. Utilized statistical tools, including Chi‑square and K‑means clustering- Providing internal tooling to reduce traction of getting new content from ideation to delivery. Such as setting up a NLP pipeline for semantic similarity on the transcription based on question context

Ahmedabad, Gujarat, India

- Participated in the initial phase of building a comprehensive ERP solution, implementing both backend and frontend from scratch.- Developed robust and scalable APIs and services for processing user data asynchronously. These contributions helped transform the application to have a more real‑time nature.- Worked on designing and developing reusable components for a Single Page Application (SPA) using React. These components ensured seamless integration with backend services and significantly enhanced user interface efficiency and overall user experience.