Developed an AI-powered voice agent for a real estate client, streamlining the sales pipeline and handling customer inquiries, containerized with Docker and deployed via CI/CD pipelines using GitHub Actions.

• Integrated real-time speech-to-text, natural language processing, and text-to-speech using Twilio, Deepgram, and OpenAI via sockets and webhooks, enabling the system to process 100+ customer inquiries daily.
• Worked on a power management system for Omron using NestJS to create a structured backend of models, controllers, and services.
• Developed and tested secure RESTful APIs with JWT authentication, supporting CRUD operations for chart data and user management.
• Built real-time dashboards in React.js and TypeScript for key power metrics, improving operational visibility and response time by over 20%.
• Implemented automated scheduling with cron jobs to generate CSV/PDF reports, reducing manual effort by 80%.
• Worked on a sports activity tracking app using Flask and React.js, implementing real-time route tracking with WebSockets and Google Maps API, improving mapping accuracy by 30%.

Ensured reliability of contract management flows used by clients like Airbnb and Apollo.

• Built and tested 20+ workflow validations for .docx templates using JavaScript, reducing JIRA-reported on-call issues by 40%.
• Created backend validators using Django to handle deleted or deactivated approvers, deployed on AWS, reducing stale contract incidents by 75%.
• Developed validations for missing or inactive default signatories to improve contract flow efficiency.
• Wrote recursive functions to traverse AST objects and verify template variables.
• Collaborated with frontend and backend teams to ensure seamless integration of all validators.

Analyzed trending keywords in medical prescriptions using Python frameworks like NumPy and Pandas to extract actionable insights.

• Developed an error correction algorithm improving medical prescription accuracy by 90% using PyTorch models for sequence prediction.
• Automated extraction of financial data from 1000+ XML files, reducing preprocessing time by 20%.
• Scraped user data from a client website using Selenium and generated CSV files to support alerts in a Django application.

Developed an innovative proof-of-concept compiler for an edtech startup product, scaling to 2000+ concurrent users using Angular, Node.js, and MongoDB.

• Contributed to codebase migration from C# to Node.js.
• Integrated support for C++, Java, and Python by creating REST APIs for compilation.
• Incorporated AI-powered debugging using OpenAI API to provide ChatGPT-based support.
• Reduced error rates in live SQL/DBMS exams by scripting automated verification of database and table creation from CSV data.
• Implemented autosave feature to prevent loss of user code during editing sessions.