• Owned and resolved 16 production incidents (P1/P2) affecting 1,000+ users; maintained availability for executive demos and critical business operations
• Completed 2,100+ oncall hours across 24/7 and business-hours rotations for AWS tooling and production infrastructure
• Mitigated packet loss incidents and recovered Kubernetes control plane by remediating failed nodes during network outages
• Coordinated cross-functional incident response with networking, security, and infra teams to minimize MTTR during critical outages

• Built SLO/SLI dashboards tracking HTTP, ICMP, SSH health, connection latency, packet loss, and throughput across 8+ global data centers
• Integrated time-series metrics for real-time service health visualization with historical trend analysis
• Tuned alerting thresholds and severity classifications to reduce false positives; improved signal-to-noise ratio for oncall alerts
• Built network probe heatmap visualizations with async data fetching from time-series databases

• Built Python pipelines processing hardware inventory across 10,000+ devices — extracting and categorizing CPU/GPU/memory specs from distributed databases
• Built 43+ SQL analytics queries for hardware resource analysis and device fault diagnostics across petabyte-scale datasets
• Authored 24 Jupyter-style notebooks for automated device categorization, network monitoring, and data quality analysis
• Migrated service APIs from legacy Thrift to thrift-python, enhancing parameter handling for telemetry and inventory updates

• Managed Kubernetes cluster operations: control plane recovery, node lifecycle management, failover handling during network incidents
• Deployed DMZ architecture at London and Ashburn data centers with reverse proxies, F5 load balancers, and MetalLB endpoints
• Automated Linux server configurations via Chef — sysctl tuning, cron job scheduling, and log forwarding pipelines
• Created VLAN validation automation tools for configuration verification via MAC address, hostname, and serial number lookups

• Implemented CapturyReplay motion capture software and hardware in Meta Workrooms and Horizons VR platform; tracked and matched Captury avatar movement, translating the Meta rig to use the same parent with existing third-party rig support
• Launched Azure infrastructure containing Azure OpenAI endpoints enabling Meta researchers to access LLM tooling within the research environment

• Architected modular "Studio in the Cloud" AWS infrastructure using Terraform, delivering reusable deployment modules for VFX/game studios running Autodesk, Unreal Engine, Nuke, and Blender workloads — for customers including Epic Games and Deck Nine Games
• Built full-stack studio management portal with FastAPI and React, featuring WebSocket real-time workstation state backed by Lambda, DynamoDB, Cognito, API Gateway, and Step Functions
• Engineered one-day studio deployment system using Python/Luigi task orchestration, Go REST API, and React intake form to provision complete cloud studios from a single form submission
• Developed Python/Slack Bolt bot for Epic Games integrating Shotgrid production asset tracking into render pipeline team workflows, deployed on ECS Fargate via Terraform and GitHub Actions CI/CD
• Designed high-availability Deadline render farm infrastructure with auto-scaling EC2 behind NLB, EFS storage, and DocumentDB for Epic Games art pipeline with AWS Portal spot instance integration
• Created and published AWS Marketplace sales demo featuring one-click EC2 workstation launcher — Go backend, Next.js/TypeScript frontend, Packer-built Windows AMIs with Unreal Engine 5 integration
• Built serverless multiplayer game backend at internal hackathon using GameLift FlexMatch for skill and latency-based matchmaking, Lambda/API Gateway for session brokering, DynamoDB for player state, and Fargate/ECR bot clients for load testing — supporting Unity, C++, and Unreal Engine clients
• Wrote Python/boto3 governance tooling to audit untagged EC2 instances and EBS snapshots across AWS regions using CloudTrail attribution

• Developed TC3 Sim medical training simulation integrating HaptX, Leap Motion, and Ultrahaptics haptic hardware
• Built WebGL-based multiplayer training environment for VA Medical Center

• Developed P8A Poseidon aircraft maintenance simulation with regression recording capabilities

• Created T-6A and IAF Air Force cockpit instruments — gauges, dials, and interactive switches — using C++

• Converted legacy weapon test application from Visual Studio 2005 to 2013; implemented TCP/IP and DDS XML messaging agents in C++