Headline

Software Engineer @ Oppex AI | Ex - EY, Nagarro

About

• Software Engineer at Oppex AI, owning end-to-end features for an AI-driven incident management platform. • Translate real customer problems into scalable system designs and implement them across the full engineering lifecycle backend APIs, data modeling, frontend integration, and deployment. • Actively contributes to improving reliability, performance, and usability through practical AI-based solutions. • Previously worked at Ernst & Young, gaining experience with large-scale enterprise systems and consulting-led problem solving. • I started my career at Nagarro, building strong software engineering fundamentals in a global MNC environment. Currently deepening expertise in backend engineering, system design, and AI engineering concepts, focused on building scalable, real-world systems.

Location

India

AI stack

claude code

claude

cursor

gemini

github copilot

mcp

Work experience

Software Engineer at Oppex AI, Jul 2025 - Present · 1 y 4 mo, Bengaluru, Backend Engineer at Oppex, building scalable incident management, on-call scheduling, and PagerDuty-to-Oppex migration systems. • 𝗣𝗮𝗴𝗲𝗿𝗗𝘂𝘁𝘆 𝗠𝗶𝗴𝗿𝗮𝘁𝗶𝗼𝗻: Migrated enterprise customers from PagerDuty and built daily reconciliation across 13,000+ incidents/week, achieving ~82% match rate. Improved title matching, reducing unmatched incidents by 54% to enable confident customer cutovers. • 𝗔𝗜 𝗢𝗻-𝗖𝗮𝗹𝗹 𝗣𝗿𝗲𝗽: Built personalized pre-shift briefs featuring predicted incidents and the top 3 likely causes linked to similar historical incidents. Integrated TypeSafe’s Jev model to validate causes and lower confidence when uncertain. Added workspace-level controls and a platform-wide kill switch. Reduced LLM costs by sharing calls across briefs with the same predicted problem and reusing results when evidence remained unchanged. • 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲: Made incident listing 3.9× faster (2.1s → 0.5s) for workspaces with 1M+ incidents through query optimization and indexing. Added a 1M-incident benchmark to validate future changes. • 𝗢𝗻-𝗖𝗮𝗹𝗹 𝗦𝗰𝗵𝗲𝗱𝘂𝗹𝗶𝗻𝗴: Built rotations, leave and shift-swap overrides, roster change history, and timezone support. • 𝗕𝘂𝗹𝗸 𝗢𝗽𝗲𝗿𝗮𝘁𝗶𝗼𝗻𝘀: Implemented role-aware bulk Resolve, Acknowledge, and Snooze. Reduced database queries for large incident exports from ~500 to 1. • 𝗣𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻 𝗘𝗿𝗿𝗼𝗿 𝗗𝗲𝘁𝗲𝗰𝘁𝗶𝗼𝗻: Built OnCallLogger to scan AWS S3 production logs and automatically raise incidents for backend errors. • 𝗔𝗹𝗲𝗿𝘁 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁: Added self-service alert grouping at rule and workspace levels, plus Slack and email notifications for individual, bulk, and automatic resolutions. • 𝗧𝗲𝘀𝘁𝗶𝗻𝗴 & 𝗥𝗲𝗹𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆: Built automated end-to-end and performance test suites, including a 1M-incident load test integrated into CI.

Senior Analyst at EY, May 2024 - Jul 2025 · 1 yr 3 mos, Noida, • Architected and maintained scalable microservices using Python, FastAPI, and MySQL for EY’s Generative AI Studio, reducing manual SDLC effort by 40%. • Designed and implemented an AI-driven pipeline leveraging GPT-4o and Azure Cognitive Search to parse complex business logic, vectorize inputs and outputs, and retrieve historical results, improving BRD and test case generation accuracy by 23% while reducing manual effort by 60%. • Developed a real-time Control Tower Logistics (CTL) backend using Spring Boot and engineered anomaly detection logic for in-transit shipments, reducing incident response time by 11%. • Implemented SAML-based Single Sign-On (SSO) integration with Azure AD across internal logistics and AI platforms, enhancing security and reducing authentication failures by 25%.

Trainee at Nagarro, Mar 2023 - Oct 2023 · 8 mos, Remote, • Built and optimized scalable backend APIs using Spring Boot and PostgreSQL, improving application performance, reliability, and maintainability. • Integrated RESTful APIs with frontend applications, enabling seamless data exchange and enhancing user experience across multiple client projects. • Optimized application state management using TypeScript and Redux Toolkit, streamlining data flow and improving code maintainability. • Reduced unnecessary component re-renders through efficient state management and rendering strategies, significantly enhancing application responsiveness and scalability.

Education

Bachelor of Technology, Information Technology, Dr. A.P.J. Abdul Kalam Technical University (AKTU), Lucknow

Public links

github: https://github.com/harsh1000-hub

leetcode: https://leetcode.com/user-19

harsh-vardhan-portfolio.vercel.app: https://harsh-vardhan-portfolio.vercel.app/