OLT/SDN Test Automation · Platform Engineering · Applied AI
Applied R&D Engineer at Nokia Networks. I drive automated regression for Lightspan OLTs and Altiplano SDN, run the team's operational platform, and ship applied AI across 8 formal initiatives.
Test automation, platform engineering, and applied AI — three parallel tracks, all running in production.
Owns 17 Robot Framework test suites (48 test cases) running as weekly regression against Lightspan OLTs and Altiplano SDN. Covers the full secure software installation path — EST mTLS, SFTP, and PKI configured certificates — on both simplex and redundant NT hardware. Every failure mode is a named test case; every teardown is deterministic so the next batch starts clean.
Impact: Over 30 defect fixes across 75 changesets. Weekly regression covering releases 2023 through 2026.
Four-module internal platform: Equipment Dashboard (OLT reservations with defined timelines and public Community Wall reminders), Network Control Plane (browser-based NETCONF/ConfD with 4 branches and pre-flight connectivity checks), KnowMore (on-premises Agentic AI), and Lightspan Health (twice-daily infrastructure monitoring).
Impact: 440+ reservation cycles. 600+ NETCONF device operations in six months. ~90% reduction in coordination overhead.
KnowMore — air-gapped Agentic AI platform with domain-specific agents powered by hybrid FAISS + BM25 + RRF retrieval and a local LLaMA 3.1 8B fallback. Zero cloud dependency.
8 AI initiatives shipped in a formal register: AI-SDLC workshop (agentic code generation into production code review), Cursor Agent Skills for test generation, wiki authoring, implementation patches and defect triage, an Efficiency Tracker mining AI session history for management reports, and two Enterprise LLM domain agents.
Nokia StrongHer Ally Award — Recognition for technical
contributions.
Developed Robot Framework test suites for Nokia Lightspan OLT (LT and NT) robustness testing. Built solution-level test cases for OLT and SDN environments. Acquired hands-on proficiency with NETCONF device management and automated regression workflows.
Integrated B.Tech + M.Tech. Graduated Rank 4 in the department with a CGPA 9.48/10. Two-time VIT Meritorious Award recipient.
On-premises Agentic AI platform. Administrators create domain-specific agents grounded in enterprise documentation — citation-backed responses, no hallucination on domain queries, fully air-gapped.
Hybrid retrieval: FAISS + BM25 fused with Reciprocal Rank Fusion (RRF). Local LLaMA 3.1 8B fallback on standard CPU — zero cloud or GPU dependency.
Browser-based NETCONF/ConfD console for Lightspan OLTs. Pre-flight TCP and NETCONF connectivity checks before any operation, asynchronous Celery execution, and a full session audit log downloadable after each run.
Impact: 600+ device operations in six months. Eliminated manual CLI dependency for routine OLT management.
Equipment reservation platform for shared IP-addressable infrastructure. Replaces ad-hoc coordination with a structured booking lifecycle — named owners, defined timelines, and automated SMTP reminders for overdue returns.
Impact: 440+ reservation cycles across 30+ chassis. ~90% reduction in coordination overhead.
Patented a real-time IoT safety solution for toddler tracking with configurable parent controls and network status visibility. Patent ID: 202241031906.
Patented an IoT leg-band wearable that detects and monitors erythromelalgia attacks using machine learning on biometric temperature and RGB colour shifts. Patent ID: 202341078712.
Published at ICSCSS 2023. Autonomous IoT inventory system with shelf-life tracking and automated stakeholder alerts. DOI: 10.1109/ICSCSS57650.2023.10169313. Over 120 full-text views.
