Tech Mahindra · Client: Zebra Technologies · Bangalore, India
Domain: Android Framework / Enterprise Android Devices
Android Recovery LED Notification Support
- Developed and deployed LED-based notification support in Android Recovery Mode using C/C++, enabling users and technicians to identify OTA update success/failure states without display dependency.
- Implemented OTA status indication by invoking LED APIs during recovery update workflows, covering edge cases including update failure, successful completion, and recovery-state transitions.
Serial Number-Based OTA Validation
- Designed and implemented serial number and device property validation logic to restrict OS updates to authorized customer-specific devices.
- Enabled validation across OTA, A/B mode, and recovery mode update flows using configuration-based checks, improving device security and update governance.
AI-Powered Android Log Analysis Automation
- Built a GenAI-powered Android log analysis system using Python, LangGraph, OpenAI/Claude APIs, Jira API, chunking, regex parsing, keyword filtering, and summarization.
- Automated analysis of logcat, bugreports, kernel logs, tombstones, ANR/crash logs, battery logs, audio logs, and NFC logs — with average log sizes of 200MB.
- Designed multi-node LangGraph agentic workflows: log summarizer, RCA agent, error classifier, Jira analyzer, similar issue finder, and report generator.
- Integrated Jira APIs to analyze 250+ existing and new issues, generate summaries, classify errors, suggest fixes, detect similar issues, and prepare Jira-ready comments.
- Used LangGraph state management, memory, and conditional routing to automate troubleshooting across Android framework and component teams.
- Reduced issue investigation time from 10–15 hours to 1–2 hours (~90% reduction), measurably improving debugging efficiency and accelerating root-cause analysis.
- Collaborated with Framework, Audio, Battery, NFC, and Stability teams to improve issue triage and speed up defect resolution.