Work Experience

• Led the end-to-end design and development of a scalable backend platform for a fitness event ticketing system, powering multi-activity events, team-based registrations, and approval-driven onboarding workflows, enabling organizers to seamlessly manage complex event structures.

• Engineered a secure and reliable payment ecosystem using Razorpay, with built-in support for coupons, discounts, ensuring accurate handling of high-value, revenue-critical transactions.

• Successfully supported over 3,500+ bookings during India’s largest fitness festival (peakst8 Festival) by building fault-tolerant systems, optimizing database queries, and maintaining high availability under extreme traffic conditions.

• Enabled direct frontend uploads to AWS S3 via pre-signed URLs, significantly reducing backend processing overhead and improving upload speeds.

• Improved platform performance and long-term scalability by implementing efficient page-based pagination, optimized API response patterns, and structured data access layers for high-traffic endpoints.

• Collaborated closely with product, operations, and marketing teams to translate business requirements into robust technical solutions, ensuring rapid feature delivery without compromising system reliability.

• Built and maintained a responsive sports community platform using Next.js, TailwindCSS, and Firebase, serving 3,500+ users.

• Engineered a Ninja Leaderboard system with streak-based scoring, and bonus point calculations.

• Designed a P2P pickleball rating system with WhatsApp-triggered notifications and a leaderboard tracking player ratings and reviewer counts.

• Evaluated and benchmarked edge computing boards by analyzing performance with and without NPU acceleration for specific workloads.

• Implemented anomaly detection on industrial weld datasets using machine learning techniques such as Isolation Forest.

• Curated and processed large-scale datasets from PDF documents containing complex, multi-layered tabular structures for downstream analysis and reporting.

• Automated text and table extraction pipelines using Table Transformers, Camelot-py, and PaddleOCR, converting unstructured data into structured JSON and Markdown formats, reducing manual processing time by nearly two weeks.

• Implemented validation and data-cleaning workflows to refine extracted outputs, ensuring high accuracy, consistency, and reliability across datasets.