Amol Pardeshi

20 NOVEMBER 2026

BENGALURU

( SPEAKER )

Senior Software Engineer

Gojek

Amol Pardeshi is a Senior Android Engineer working at Gojek as a part of its Mobile Platform Team. Gojek is the largest super app in South East Asia boasting 100M+ users. Amol has over 6+ years of experience in Android development and for the past three years, he has been focused specifically on Android Performance Engineering.
His contributions include developing a custom ANR SDK from scratch for Gojek based on AppExitInfo, watchdog, and native SIGQUIT signal handling with proactive regression alerts. This along with other performance improvements including optimization of Dagger graphs, fixing of background processes, and implementation of AI-driven code guardrails led to a reduction of Gojek’s user perceived ANR rate by 80%. His areas of expertise include crash rate control, memory leaks avoidance, app size minimization, and app startup optimization with large multi-module Kotlin codebase.
Amol has spoken at droidcon India 2025 on Gojek's ANR reduction path and is an organizer in Bangalore Android community (BLRDroid).

Curated ANR and Leak patterns to avoid - https://mintlify.wiki/AmolPardeshi99/android-performance-skills
Agentic skills to write ANRs and Leaks free code -
https://www.skills.sh/amolpardeshi99/android-performance-skills/android-performance-anr-and-jank-prevention
https://www.skills.sh/amolpardeshi99/android-performance-skills/android-performance-memory-leak-prevention
ANRLab – ANR Tooling: https://github.com/AmolPardeshi99/ANRLab/

Lightning talk

The ANRs You Can't See: Gojek's 80% Reduction with Custom Detection and AI Guardrails

Gojek, the super app of Southeast Asia with 100M+ downloads, has reduced their ANR rate from 1.3% to below 0.3%. The easy bit was resolving main thread ANRs visible in user feedback reports. Here I'm going to talk about the hard part, ANRs not visible in Play Console, ANRs in background processes, and regressions your team keeps delivering. Here I will take you through three levels of our stack: 1. Custom ANR detection: AppExitInfo, main thread watchdog, and native SIGQUIT handling driving a dashboard with pre-ANR/post-ANR session context 2. Background ANR mitigation: FCM process migration, splitting a god Dagger component with 500+ bindings serializing, custom high frequency view optimization 3. AI in the loop to prevent ANRs/Leaks agent skills and context files blocking the ANR/Leak inducing code generation, automated LeakCanary to MR Pipeline 100% actionable patterns and agent skills you can leverage directly here: https://mintlify.wiki/AmolPardeshi99/android-performance-skills