Avi Narayan

Platform Architecture & Scalability Leadership
Owned the end-to-end architectural evolution of high-traffic NCX backend systems, driving platform-wide scalability, reliability, and performance. Led critical redesign initiatives to eliminate systemic bottlenecks and enable sustainable growth across services handling millions of requests.

Distributed Systems Performance Optimization
Drove deep performance optimizations across core data processing pipelines, including ConversationActivitySummary, by re-architecting execution flows and eliminating redundant computation layers. Delivered ~35% latency reduction and 2.1x throughput improvement, significantly enhancing system efficiency at scale.

Cloud-Native Service Design (Media & Asset Platform)
Designed and delivered a multi-tenant, cloud-native asset management platform for logo uploads and retrieval. Established highly resilient, low-latency pipelines using pre-signed URLs, intelligent caching, and distributed fault-handling mechanisms, achieving ~99.98% success rates and sub-linear latency scaling.

Reliability Engineering & Fault-Tolerant Systems
Institutionalized reliability-first design principles across services, introducing retries, circuit breakers, and graceful degradation strategies. Enabled ~99.99% availability with zero-downtime rollouts, ensuring consistent performance under peak traffic and partial system failures.

API Platform Modernization & Technical Debt Elimination
Led modernization of critical integration APIs, aggressively reducing legacy surface area (~75%) while improving correctness and maintainability. Optimized system latency (p99 ↓ ~28%) and established cleaner contracts to support long-term extensibility and cross-team adoption.

Analytics Platform Re-Architecture
Re-architected analytics systems by enforcing strict separation of concerns between orchestration and connector layers. Centralized business logic into reusable utilities, reducing system complexity (~45%) and enabling faster feature iteration without increasing code footprint.

Real-Time Data Processing & Observability
Engineered concurrent data aggregation pipelines to support real-time analytics use cases, including rolling KPIs and trend analysis. Reduced response latency from 1.8s → 650ms (~64%), enabling near real-time decision-making capabilities for downstream consumers.

AI Workflow Enablement & Backend Strategy
Led backend design for AI-powered product workflows (e.g., “Create Site”), defining secure and efficient service-to-service communication patterns. Reduced authentication overhead (~30%) using JWT-based models while enabling scalable automation pipelines.

Identity & Authentication System Simplification
Redefined SSO architecture by shifting to JWT-derived tenant identity, eliminating configuration-heavy flows and reducing production incidents by >90%. Maintained full backward compatibility while significantly simplifying developer and client integration experience.

Data Integrity & Metric Consistency
Designed idempotent aggregation pipelines for critical business metrics (e.g., lifetime unique visitors), ensuring 100% data consistency across distributed systems. Established reliable data contracts for analytics and reporting layers.

Backend Integration for New Business Line
Architected and implemented the complete backend integration for a new Swiggy business line focused on personalized offers and discounting. Designed dynamic workflows for item, customer, and multi-store level offers.

Scalable Offer Evaluation System
Built scalable and configurable pipelines for offer creation and evaluation using tenant-specific SLA logic, discount constraints, and metadata resolution, enabling rapid expansion across verticals.

Enhanced SLA & Metadata Transformations
Integrated new delivery promise types, SLA strategies, and gratification models for accurate delivery-time guarantees and consistent metadata, used across ordering and fulfillment platforms.

Test Automation & Validation Frameworks
Designed reusable evaluation test cases and mocks across business lines, reducing test flakiness and improving CI reliability. Increased release confidence and reduced regression effort by over 60%.

Platform Configuration & Limits Tuning
Scaled offer platform by implementing constraint evaluation logic and increasing configurable limits (e.g., max offers per user per SPIN) for operational flexibility.

API Design & Cross-Functional Collaboration
Shipped reliable, low-latency APIs for homepage, cart, and checkout integrations. Worked with product, frontend, and data teams to ensure personalized offer delivery and seamless user experience.

Business Line Expansion Across Vertical Stack
Enabled consistent offer behavior and SLA guarantees across Swiggy’s verticals like Instamart, Dineout, Genie, and the newly launched Assure line using a common transformation and evaluation backbone.

UPI Payment Gateway Integration
Architected and implemented the backend integration of the UPI payment gateway, enabling real-time payment processing for millions of users during sales and regular checkout flows.

Scalable Microservices for Payments
Developed containerized microservices architecture using Docker and Kubernetes, ensuring fault-tolerant and auto-scalable payment processing infrastructure.

Secure & Compliant Payment Workflows
Integrated advanced encryption (AES) and public-key infrastructure (PKI) into payment workflows to protect sensitive user data and meet industry-grade compliance requirements.

Database Optimization for Transaction Logs
Optimized SQL queries and introduced smart indexing to accelerate transaction log retrieval by 40% during peak loads, significantly improving payment throughput and debugging capabilities.

Real-time Payment UX Enhancements
Collaborated with frontend teams to deliver low-latency APIs for real-time transaction updates, providing users with immediate payment status and enhancing trust during high-stakes checkouts.

High-Throughput Event Processing with Kafka
Led backend Kafka integration for payment event processing and notifications, enabling efficient scaling of event streams during high-traffic events like Big Billion Days.

Conversational Bot – Video Assist Enhancements
Designed enhanced user flows for the Video Assist Bot, introducing user consent protocols, Redis-based state management, and auto-transfer logic, reducing wait time and improving session continuity.

Live Assist Agent Availability & Experience
Upgraded backend logic for Live Assist to dynamically manage agent availability and improve queueing, boosting feature adoption during key sale periods.

Flipkart Smart Upgrade Plan – Routing Migration
Contributed to platform migration from legacy routing to a new load-balancing algorithm, improving reliability and response times for the Smart Upgrade feature.

Issue Resolution & UI Collaboration
Troubleshot system downtimes and collaborated with UI teams to improve Live Assist experience, especially under sale-induced high load.

• Enhanced interfaces to promote better functionality for users.
• Researched and adopted new technologies to add value to existing offerings.
• Developed technical requirement documentation for new software products.
• Improved systems with addition of new features and infrastructure.
• Investigated system issues and implemented resolutions to reduce downtime.
• Created new API for unread message count of chat folders, groups from specific timestamp to current timestamp.
• Worked on Event-Driven state machine implementation of messaging store for Voicemails.
• Worked on various API flows like Provision, Deposit, Notification, Folder operations for message store ( Used C++, Cassandra Prepared Statements)
• Worked on Key Performance Indicator(KPI) and critical data store (TRL) during runtime.
• Worked on Db Latency on HTTP request of message store.

• Created and optimized records management strategies to coordinate and protect information.
• Produced high-quality documents, spreadsheets and presentations for internal and customer-facing needs.
• Analyzed problems, identified solutions and made decisions.
• Maximized productivity by analyzing protocols and identifying areas for improvement.
• Completed research, compiled data, updated spreadsheets and produced timely reports