Software Engineer — Cisco Systems Pvt. Ltd.

MAC-Segmentation & Workflow Modernization

Aug 2024 – Feb 2026

Objective

Make switch-level (Layer-2) traffic visible using Security Group ACL logs to enable auditability and faster troubleshooting.

Approach

• Implemented comprehensive Layer-2 flow logging in SGACL with structured tagging to ensure device-to-device traceability

• Traced SGACL and ACLLOG execution paths end-to-end to identify and fix silent logging failures

• Ensured logs were accurate, consistent, and compliance-ready for forensic analysis

Impact

• Delivered audit-ready Layer-2 flow trails showing exact communication paths, timing, and participating endpoints

Objective

Stabilize long-running debug sessions and eliminate crashes impacting investigation workflows.

Approach

• Identified and patched a critical ACLLOG memory leak using core-dump analysis and system-log correlation

• Removed crash conditions blocking extended debugging and customer investigations

Impact

• Significantly improved logging system stability, enabling reliable long-duration investigation sessions

Objective

Reduce bug triage time by augmenting engineers with AI-assisted evidence retrieval.

Approach

• Built a retrieval-augmented generation (RAG) assistant querying historical bugs, support tickets, and system logs

• Designed automated correlation pipelines to surface actionable root-cause hypotheses

Impact

• Reduced root-cause analysis time by approximately 60%, accelerating defect resolution and team throughput

Objective

Modernize NX-OS configuration workflows by migrating from legacy PSS to the DME object store.

Approach

• Used GitHub Copilot and LLM-assisted prompting to convert legacy PSS session data into DME-compatible configurations

• Standardized and automated the migration workflow to ensure repeatable, low-risk outcomes

Impact

• Automated repetitive configuration tasks, reducing manual effort and improving platform consistency

• End-to-end tracing is critical in distributed systems — silent logging failures can mask high-impact issues

• AI tooling (LLMs, RAG) accelerates knowledge-heavy workflows but still requires human validation

• Stability fixes and memory-leak remediation often deliver outsized user impact

• Structured logging and metadata are foundational for trustworthy production audit trails

Networking: Layer-2 switching, SGACL, ACLLOG, NX-OS internals
Debugging: Core dumps, system-log analysis, memory profiling, root-cause investigation
AI/ML: Retrieval-augmented generation (RAG), LLM integration, prompt engineering, GitHub Copilot
Automation: Configuration migration pipelines, workflow automation
Software Engineering: Systems design, debugging at scale, cross-team collaboration