Muhammad Usman

Software Engineer III

Cisco Systems • August 2020 – Present

Bangalore, India

• Architecting a unified enterprise portal for reverse logistics operations, building AI agents powered by LLMs to automate failure analysis workflows—projected to serve 500+ users across 10 global manufacturing sites, improving data quality and reducing case resolution time by 20-30%.
• Built conversational AI chatbot for ML-powered PCB fault diagnosis using RAG architecture and vector embeddings, achieving 50% reduction in repair time across pilot factories.
• Designed and implemented Knowledge-Based Search platform using Elasticsearch and semantic search, unifying 15+ data sources including Service Requests, RMAs, and failure reports, reducing failure analysis time by 30%.
• Led migration of 10+ legacy monolithic applications to cloud-native Kubernetes microservices architecture, implementing service mesh, auto-scaling, and CI/CD pipelines, improving system reliability to 99.9% uptime and enabling 10× load handling.
• Mentoring 3 junior engineers on best practices in system design, code reviews, and agile methodologies while collaborating with cross-functional teams across US, China, and India.

Technical Undergraduate Intern

Cisco Systems • January 2020 – June 2020

Bangalore, India

• Developed real-time Grafana dashboards monitoring CI/CD pipeline health and tool adoption metrics across 15+ hardware manufacturing teams, driving adoption from 60% to 100%.
• Implemented automated deployment gates and validation checks in Jenkins pipelines, eliminating manual deployment errors and ensuring consistent hardware testing environments.
• Created comprehensive user documentation with Sphinx to streamline onboarding and enable independent usage.

Google Summer of Code

Wikimedia Foundation • May 2019 – August 2019

• Improved Wikipedia's cross-language article recommendation pipeline by integrating user engagement metrics, achieving 50% latency reduction.
• Contributed to 2× increase in editor engagement and 3.2× increase in article creation rates through enhanced recommendation quality.