Selected work
A non-exhaustive list of things I'm proud of.
2026
Leading a phased migration of critical rack-level firmware components from C/C++ to Rust to improve memory safety, security posture, and long-term maintainability. Driving the strategy, owning the POCs (notably daemon and control-plane paths), coordinating cross-team reviews, and shaping how AI-assisted tooling accelerates and validates the migration at scale.
2026
Owning and executing a multi-phase effort to strengthen the security baseline of rack management components through threat-model reviews, mitigation of high-priority findings, and concrete hardening features. Reduces risk across production fleets and next-gen platforms, with clear accountability for driving items to closure and aligning security outcomes with semester commitments.
2026
Driving a data-backed reliability workstream focused on improving rack and controller availability by identifying systemic fault patterns, improving recovery workflows, and closing automation gaps. Leading discussions, publishing availability insights, and influencing both firmware changes and operational processes to reduce MTTR and prevent repeat incidents.
2025
Lead the development of software and firmware solutions for the ARM-based Rack Manager — overseeing servers, BMCs, accelerators, PDUs, and GPUs — operating on 400K+ devices and 4M+ nodes, enabling fleet-wide monitoring, lifecycle management, and operations across Azure datacenters.
2025
Currently playing a key role in GBx00 NVL72 deployment, supporting GPU cluster buildout and enabling GPU telemetry, firmware updates, diagnostics, and Azure fabric APIs via rack-level device services.
2024
Built an AI-powered Firmware Log Analyzer Copilot that ingests logs and test artifacts, auto-detects anomalies, and generates debugging insights — replacing manual multi-hour triage workflows.
2024
To accelerate team productivity, built internal AI chatbots and championed Copilot adoption, achieving 100% usage across the team.
2023
Architected redundant telemetry systems (OOB, in-band, and serial fallback) to maintain observability during network and platform failures.
2024
Built AI-assisted RCA workflows correlating telemetry, logs, and incidents to accelerate debugging and reduce repetitive analysis.
2023
Led multi-phase security initiatives (AzSecurity) including threat modeling, risk tracking, and feature enablement across RM/RSCM.
2023
Enhanced firmware security through at-scale syslog monitoring, rate limiting, privilege separation, and Platform Root of Trust.
2024
Built and scaled telemetry pipelines and KQL dashboards for real-time analytics targeting GPU, liquid cooling, and NVL72 rack telemetry — improving device availability and enabling proactive alerting for critical scenarios.
2022
Acquired hands-on experience in board bring-up, platform initialization, board support packages, security, attestation, telemetry pipelines, and distributed cloud services for Azure AI supercomputers.
2020
Developed platform software for IOS-XR network OS, supporting large-scale routing and switching systems.
2019
Designed hardware abstraction APIs and data models, improving platform modularity and portability.
2017
University of Southern California (USC).