Commerce & Quick-CommerceEngineering Manager (Managed 25+ engineers across 6 products)

Scaling Dark Store Expansion and Quick-Commerce Fulfillment

How an automated expansion platform cut store setup time from 160 to 85 days across a 500-store national rollout.

COMPANYBlinkit(Part of Eternal, formerly Zomato)
LEADERSHIP ROLEEngineering Manager (Managed 25+ engineers across 6 products)
TIMELINENov 2022 – Jan 2026
LOCATIONGurugram / Noida, India

01 The Business Situation

During India’s quick-commerce hyper-growth surge, Blinkit was scaling from hundreds of dark stores to over 500 fulfillment hubs nationwide. Rapid rollout velocity was the single biggest commercial advantage.

02 What Was Broken

Store expansion was manual: site procurement, equipment provisioning, catalog seeding, and network setup took up to 160 days per store.

Backend Golang microservices lacked uniform analytics, graceful shutdowns, and unified linters, causing telemetry drift across teams.

Mobile app releases were fragmented across disparate Android and iOS teams, slowing weekly release cycles.

Central data reporting suffered from ETL lag across dark store inventory and order allocation systems.

03 What I Decided & Why

DECISION 1

Architected the Locus dark store expansion platform

Architectural Rationale:

Automated site onboarding, network provisioning, and dark store readiness checklists into a unified operational workflow, cutting setup time by 47% (160 to 85 days).

DECISION 2

Built an enterprise-wide Golang service kit

Architectural Rationale:

Standardized graceful shutdowns, Datadog tracing, structured logging, and reusable HTTP/gRPC middleware, cutting new microservice launch time by 40%.

DECISION 3

Spearheaded Kotlin Multiplatform (KMP) + Compose migration

Architectural Rationale:

Consolidated domain, network, and storage code into a single shared repository, achieving 80%+ code reuse across Android and iOS and cutting release cycles by 70%.

DECISION 4

Built automated Apache Airflow data pipelines and Pinot virtual tables

Architectural Rationale:

Replaced manual reports with real-time operational dashboards for store ops, franchise onboarding, and on-call alerting.

04 Measurable Results

160 → 85 Days (47% cut)Store Setup VelocitySource: Measured across Locus platform deployment during 500-store rollout
80%+ Shared CodeMobile Code ReuseSource: Kotlin Multiplatform enterprise migration across Android and iOS
40% ReductionNew Service Launch TimeSource: Golang company-wide microservices kit adoption
40% ReductionFranchise Partner ChurnSource: AI-assisted lead-scoring and partner-matching onboarding platform

05 Lessons for Startup Founders

  • Do not maintain parallel mobile codebases if domain logic is identical. Kotlin Multiplatform or cross-platform code sharing saves months of executive synchronization.

  • Invest early in a shared service kit. If every engineer invents their own HTTP logger, metrics sink, and error handler, you will waste 30% of engineering time on plumbing.

  • Treat store and fleet onboarding as an automated software workflow rather than an ad-hoc operational spreadsheet.

06 Technologies Used

GolangKotlin MultiplatformJetpack ComposeApache AirflowApache PinotDatadogKubernetes

WORK WITH UMESH GUPTA

Ready to scale your architecture
without expensive trial and error?

Book a 20-minute call to evaluate your engineering bottlenecks. Response: Reply the same business day (IST).

WhatsApp (+91 9717486696)