Bharathi Sekar

Guided students in the Human-Computer Interaction course, leading exercise classes and teaching HCI concepts and Figma for mobile application design. Evaluated assignments and projects, providing constructive feedback aligned with course objectives. Offered one-on-one student guidance and clarified key HCI concepts during exercises.

Enhanced web application performance, security, and usability through targeted feature improvements. Managed databases for data extraction, transformation, and reporting. Developed and maintained backend services, automated critical batch processes for production data flows, and supported cloud migration for B2B and B2C applications. Designed and implemented APIs for system integration and data exchange, and developed front-end pages to enhance user experience.

Develop and maintained web and desktop applications, managing version control and resolving critical production issues through structured incident management processes. Diagnosed and resolved performance bottlenecks through root cause analysis, improving system reliability and efficiency. Managed large-scale batch data processing and data validation workflows for job orchestration and execution, ensuring stable and timely operations.

Completed training in full-stack development with hands-on experience in Python, Java, SQL, HTML, CSS, React and GitHub for version control. Developed a file transfer system between an insurance company and a bank to automate payment processing for users.

Master's thesis • Architected a multi-agent LLM system using the OpenAI API, with role-based prompt engineering and context-aware response generation, to study conversational inoculation against misinformation, contributing to responsible AI research on bias mitigation. • Implemented structured system prompts and JSON-based context injection to dynamically guide distinct bot personas and ensure realistic, task-oriented behavior. • Developed a rule-based multi-party turn-taking framework with urgency scoring and speaker nomination to coordinate multi-agent conversations. • Built a web platform using Python, JavaScript, and a SQL database for user experiment to support real-time interactions, session storage, and task management, and deployed the system using Docker.

Thesis: New Approach for Social Distance Detection Using OpenCV and Clickstream Approach to Detect Malicious Social Bots.