Swaraj Patil

Shipped the company's first AI-powered interview web app, building the NextJS + TypeScript interface from wireframe to production.

LLM based Multi modal Personalized Interviewer that generates role-specific questions and real-time followups, keeps candidate in conversational mode.

Designed clean REST APIs in Django for user management, interview data, & secure authentication systems, consumed by React dashboards with advanced filtering & bulk actions.

Implemented Socket.io-based video/audio recording system with WebSocket support, enabling asynchronous interview reviews & reducing screening time by 25%.

Used Django ORM for efficient database modeling with PostgreSQL, ensuring data integrity & optimized query performance.

Built Certifier — a web-based certificate issuer for course participants with dynamic PDF generation & verification system.

DevOps responsibilities: Created GCP VM instances, dockerized applications, and deployed on VM instances.

Azure OpenAI resource management: Created Azure OpenAI resources for course participants (org-level categorization), deployed LLMs (GPT-4.1, GPT-4.1-mini) & embedding models (text-embedding-3-small, text-embedding-3-large).

As assistant instructor, created complete course content & conducted three courses — AI for Engineers (GSK, Trane Technologies) and AI for PM at AngelOne.

Created a self-serve platform using React.js and Next.js, enabling user registration and PDF profile generation using Handlebars, Puppeteer, and EJS.

Implemented component-based architecture using React.js UI components with Redux Toolkit and Material UI, ensuring users only accessed data relevant to their roles, enhancing security.

Achieved a 30% reduction in page load time through optimization and caching strategies, improving performance and user experience.

Collaborated with DevOps Team to containerize the front-end and handle deployment on GCP linux servers.

Deployed AI assistants in Google Chat via Google Apps Script, streamlining collaboration and automating workflows.