Harshit Pentyala

• Diacritic Project: Enhanced booking flow by implementing logic to handle diacritic accent names, converting unsupported characters to ASCII. Developed and integrated new C++ library code with the Rail component.
• API Integration: Worked on SOAP and REST-based architectures to manage and transform booking-related data between systems.
• Bug Fixing & RCA: Performed root cause analysis and resolved production issues, including incorrect error messages and ticket exchange failures involving active and past tickets.
• Middleware Upgrades: Led version upgrades for 42 middleware components and libraries using C++ and Python, ensuring smooth system migrations every six months.
• XCM Project: Implemented discount and voucher logic by integrating RNT, RDP, and RIN components, enabling seamless display and application of offers between shopping and booking pages.
• Innovation Initiatives: Participated in hackathons and IP week, optimizing ticket management logic to prevent duplication and improve overall ticket server efficiency.

• Local File Search Engine:
  Developed a file search application using Python and Django, enabling efficient local system searches. Built an admin interface to store previous searches and reduce redundant operations, with functionality to manage and delete search history for optimized performance.
• Posmo – Machine Learning Project:
  Designed and implemented a machine learning-based solution leveraging web scraping to gather data, generate optimized titles, and create relevant business quotations for startups.
  Developed REST APIs using Django and built a dynamic Angular-based UI.
  Utilized Random Forest to train the model, achieving 94.5% accuracy.
  Integrated an SQL database for structured data storage and retrieval, improving search visibility and relevance for startups on Google.

AI/ML-Based Automatic Ticket Assignment & Notifications:
Developed a machine learning-driven system to automate ticket assignment and send real-time notifications, improving operational efficiency and reducing manual workload.