NBA Digital Operations Center
Streaming Operations Engineer. The technical bridge between frontline monitoring and specialist engineering teams — independent investigation, advanced triage and approved mitigation, on live global sport at 99.5%+ uptime.
The chain I work across
Eight stages sit between a camera in an arena and a phone in someone’s hand. Every one of them can fail in a way that looks the same to the person watching.
- 01CaptureSignal acquisition from cameras, feeds, and satellite
- 02EncodeCompression and format conversion
- 03PackageSegmentation and manifest generation
- 04OriginContent storage and initial delivery
- 05CDNGlobal edge distribution
- 06PlaybackClient-side rendering and adaptation
- 07TelemetryReal-time quality metrics
- 08ResponseIncident detection and mitigation
The work
- →Live, linear and VOD monitored end to end across web, mobile, connected TV and streaming devices
- →Real-time watch on video, audio, captions, metadata, ad signalling and playback quality
- →Event lifecycle: readiness checks, stream validation, launch support, failover, post-event verification
- →Fault isolation across contribution, encode/transcode, packaging, origin, CDN, DRM, entitlement, ad insertion and playback
- →Incident intake → impact assessment → mitigation → escalation, coordinated with engineering, production, CDN, vendor and platform teams
- →QoS/QoE: startup failures, rebuffering, playback errors, bitrate shifts, CDN performance, device-specific faults
- →Root cause from logs, dashboards, API responses and telemetry, with customer-facing issues reproduced on-device
- →Runbooks, event checklists and escalation paths; post-incident reviews feeding monitoring and automation gains
- →Apple Vision Pro QA: immersive video, spatial telemetry, tabletop and holographic modes
- →Rotational 24×7 coverage for NBA, WNBA, G League and international properties, including high-profile live events
What I built
Observability dashboards in Python, FastAPI, React and SQL that cut incident detection time by roughly 40%, plus automation across media operations, QA, publishing and live-to-VOD that raised throughput about 30% and reduced manual error around 25%.
Beyond the control room
Lead creatives and logistics for BudX NBA House— Mumbai in 2025 and Delhi in 2026 — the league’s flagship fan events in India. Different discipline entirely, same constraint: it happens live and it only happens once.
The stack
What it taught me
A fault anywhere in eight stages looks identical to the viewer, so isolation has to come before repair — and it has to happen while the event is still running.
Reliability is not a number on a dashboard; it is whether a fan in another timezone sees the shot go in.
The fastest way to shorten an incident is to make the signal legible before the incident starts, which is why the dashboards mattered more than the fixes.