Custom-built Prometheus exporter that scrapes live data from a Neptune Apex reef controller and exposes tank parameters — temperature, pH, salinity, and more — as Prometheus metrics for Grafana dashboards. Runs as a Docker container on the homelab HPE server. Listed on the official Prometheus exporters documentation.

A Cockpit plugin that brings Download Superstation — a self-hosted torrent manager backed by libtorrent — directly into the Cockpit web interface. No separate login required; Cockpit's own authentication gates access. Listed on the official Cockpit Applications page.

Custom Prometheus exporter for Schneider Electric's Data Center Expert, integrating via SOAP API to collect metrics and alerts from RPDUs, UPS, ATS, CRAC units, CDUs, NetBotz, and generators. A single instance aggregates data from 9 DCE servers covering ~27,000 devices across Qualcomm's global data centers, surfacing environment metrics in Grafana. Deployed via a Jenkins CI/CD pipeline with automated build, test, and release.

Custom Prometheus exporter that integrates with ITAdvisor's ETL database to expose data center capacity and utilization metrics, rack counts and density, rack purpose breakdowns, default KPIs, and custom-generated KPIs. Deployed across over 1,000 data centers globally, surfacing DCIM insights directly in Grafana. Built collaboratively and deployed via a Jenkins CI/CD pipeline with automated build, test, and release.

Built an API that automatically manages the list of servers used by Prometheus HTTP Service Discovery, dynamically onboarding and offboarding servers into the IPMI Exporter target list. Balanced endpoints across a federated Prometheus setup of 3 VMs and 2 hardware nodes.

Solo-built web platform and automation suite managing Qualcomm's server procurement pipeline from RFP to PO, built on Jira's API. A service account on vendor mail lists automatically intercepted communications, capturing metadata from every message, request, and quote file across thousands of RFPs per year. Powered vendor scorecards, market-wide pricing and availability trend analysis, and demand forecasting from historical data. Presented internally as 'RFQ / RFP Analytics (Vendor Performance)' at Qualcomm.

Automated tool that tracks OEM and third-party support coverage for all Qualcomm servers globally. Pulls server inventory from DCIM and CMDB, collects contract data from OEM APIs and web scraping, stores results in a database, and feeds PowerBI dashboards used for annual maintenance contract reviews.

Built the platform that manages a fleet of 1,500 servers available for planned maintenance windows, tech refreshes, and project kickoffs. Tracks utilization and performance, monitors server health, and automates configuration tasks across the fleet.

Custom web application for configuring, reconfiguring, and sanitizing data center infrastructure devices including RPDUs, PDUs, UPS, CRAC units, CDUs, and AKCP generators, as well as monitoring devices like NetBotz and RFCode. A ground-up rebuild of a tool originally developed in 2015.

Deployed a PXE boot environment that loads a minimal live Linux OS entirely into memory on servers awaiting imaging or production. Custom scripts and services self-report system information into a central repository, making compute waste visible and enabling automated tracking of unprovisioned hardware.

Solo-built QGenie-powered tool that lets users create and modify Grafana dashboards through natural language. Works in conjunction with the Grafana OSS GitHub Integration — translating user intent into dashboard JSON changes committed to GitHub, which sync back to Grafana automatically. Supports creating dashboards from scratch, adding and modifying panels, changing visualization types, and updating queries across all supported query types including PromQL and LogQL.

Solo-built background service that adds bi-directional dashboard-as-code sync between Grafana OSS and GitHub — replicating a feature otherwise exclusive to Grafana Cloud. Dashboard changes in Grafana are automatically committed to GitHub, and changes pushed to GitHub are reflected back in Grafana, enabling version control, code review, and GitOps workflows for dashboard management on self-hosted infrastructure.

GitHub Actions workflow that automatically deploys this site on every push to main. SSHes into the DigitalOcean droplet, pulls the latest code, installs dependencies, and restarts the service — zero manual steps from commit to production.

FastMCP-based MCP server connecting Qualcomm's QGenie generative AI platform to Jira. Enables the Universal Platform Integration team to manage their full project workflow through natural language — creating, editing, searching, closing, and linking Epics, Stories, and Sub-Tasks, as well as setting properties and attributes. Reduces context-switching and manual ticket management by bringing AI directly into the team's sprint workflow.

Solo-built MCP (Model Context Protocol) server integrating Qualcomm's internal generative AI platform, QGenie, with Confluence. Allows users in the Data Center & Compute Services org to interact with Confluence documentation entirely through natural language — creating, editing, deleting, searching, summarizing, merging, and deduplicating pages via AI. Eliminates manual documentation overhead and brings GenAI directly into the team's knowledge management workflow.

Role-based AI personas for QGenie forming a multi-agent code review pipeline. A project passes sequentially through four personas — Software Engineer, Peer Reviewer, IT Security & Risk, and Technical Writer — until it clears every gate. Enforces consistent quality, security, and documentation standards across projects.

Claude CLI and GitHub Copilot are integrated into daily development and infrastructure work. Used for automation scripting, code review, API design, writing Prometheus exporters, and accelerating complex problem solving across Python, Bash, and infrastructure tooling.