Mayank Agarwal

Experience

13+ years across ML platforms, search, distributed systems, and supply chain.

Principal Architect

Zepto · Bangalore

Nov 2022 – Present

Led technical strategy for platform availability, infrastructure optimization, and platform initiatives during the hypergrowth phase — scaling operations 10x while strengthening unit economics ahead of IPO readiness.

Platform & Applications

  • Architected comprehensive search infrastructure from scratch — LLM-augmented autosuggestion, query understanding (intent prediction, correction, segmentation), semantic retrieval, and cohort-aware MoE ranking — processing millions of daily queries with measurable uplifts in orders/day, AOV, and gross profit.
  • Drove a GenAI development platform empowering cross-functional teams to ship production-ready AI solutions, with built-in observability, cost tracking, and A/B testing.

Platform Modernization & Scale

  • Led migration from a monolith to microservices, improving platform availability from 94.6% to 98.7% and enabling independent service scaling.
  • Architected systems supporting 10x growth from 250K to 2.5M orders/day with higher reliability and reduced system-wide failure impact.
  • Built a real-time observability framework across distributed systems, cutting MTTR by 60% and enabling proactive issue detection.

Cost Optimization & Infrastructure

  • Spearheaded an enterprise-wide infrastructure optimization initiative, reducing cloud spend by 47% while holding performance SLAs — directly strengthening unit economics.
  • Designed cost-aware auto-scaling strategies, optimizing resource utilization across 200+ microservices.
  • Established FinOps practices and tooling for continuous cost monitoring across engineering teams.

Supply Chain & Logistics Systems

  • Designed the end-to-end supply-chain technology stack — dark-store management, mother-hub operations, and inbound/outbound logistics workflows.
  • Implemented inventory control systems for precise stock management, reducing carrying costs and waste by 35%.
  • Built a last-mile logistics optimization engine integrating route planning, rider allocation, and delivery prediction models.

Technical Leadership

  • Defined the technical roadmap and architecture standards across platform, data, and infrastructure domains.
  • Mentored 20+ engineers across teams, establishing a culture of technical excellence and operational discipline.
  • Drove adoption of modern practices — GitOps, infrastructure-as-code, automated testing, and observability-first design.

StackKubernetesAWSMicroservicesDify (Agentic AI)PythonGoOpenSearchRedisPostgreSQLMongoDBKafkaPrometheusGrafana

Engineering Lead

Navi Technologies · Bangalore

Aug 2021 – Oct 2022

Led General Insurance (GI) and CRM engineering teams, building the core insurance platform and operational systems from the ground up.

  • Built the GI core platform (product, premium, policy, payments), scaling with 99.9% uptime.
  • Architected a ticket-management system and workflow-engine platform, powering chatbot flows for automated customer query resolution and business workflow automation.
  • Fostered an observability culture across GI and CRM teams, cutting incident response time by 50%.
  • Established a framework for tech-debt identification and prioritization, improving maintainability and development velocity.
  • Led two engineering teams of 6 (GI) and 7 (CRM), owning end-to-end delivery and technical decision-making.

StackJavaSpring BootPostgreSQLKafkaRedisAWSKubernetes

Tech Lead (L4)

Gojek India · Gurgaon

Nov 2019 – Aug 2021

Led the post-ordering experience for GoFood (Indonesia, Singapore, Thailand, Vietnam), serving 1.6M daily orders.

  • Drove a 5% increase in business conversion (88% → 93%) through systematic optimization of post-order customer, merchant, and driver experiences.
  • Architected real-time order tracking and notification systems handling millions of concurrent connections.
  • Designed a fault-tolerant order state-management system with <0.1% error rate across a distributed architecture.
  • Led a team of 4 backend engineers, 2 frontend engineers, and 1 QA — responsible for hiring, coaching, and technical mentorship.

StackGoPostgreSQLRedisKafkaAWSKubernetes

Backend Engineer

Swiggy · Storefront Team, Bangalore

Jan 2018 – Nov 2019

Built recommendation and merchandising systems for the food-discovery platform.

  • Designed and implemented a Netflix-inspired collections framework enabling merchandised and personalized recommendations via context-aware collaborative filtering.
  • Contributed to the 300K daily orders milestone through improved discovery and conversion optimization.
  • Introduced Menu Items as a first-class entity, enabling business-driven assortment capabilities and dynamic item recommendations.
  • Hired and mentored engineers at SDE2 and SDE3 levels, establishing technical standards for the team.

StackJavaSpring BootElasticsearchRedisPostgreSQLAWS

Co-Founder / CTO

Help Inc. · Delaware / Bangalore

Jan 2016 – Dec 2018

Founded and led engineering for AI-powered conversational platforms.

  • Built and led a team of 7 engineers (5 dev, 2 QA), defining product roadmap and engineering execution.
  • Architected Bottr using VIPER architecture and native iOS development for personalized AI assistants.
  • Developed a knowledge-graph generation pipeline using publicly available data (Wikipedia), OpenIE, entity extraction, and POS tagging for automated response generation.
  • Built HelpTap, a chat-first P2P platform connecting users with domain experts based on query matching and expert scoring.
  • Implemented spell-correction and semantic question-matching algorithms achieving 85% response accuracy.

StackiOS (Swift)PythonNLP (OpenIE, POS tagging)Neo4jPostgreSQLAWS

Software Development Engineer

Microsoft · Hyderabad

Jul 2013 – Dec 2015

Shipped the Office suite on Android to 10M+ users and built CI/CD reliability systems.

  • K2 Project: launched the Office suite (Word, Excel, PowerPoint) on Android, owning telemetry and crash-reporting features for millions of users.
  • Designed and implemented a self-healing CI/CD system, raising build reliability from 75% to 95% and enabling faster collaborative development for globally distributed teams.
  • AudioNote: won the People's Choice Award and 2nd place in a company-wide 2-day Android hackathon for an innovative note-taking app.

StackAndroid (Java)C++AzureCI/CD tooling