Staff Technical Program Manager, Fleet Health Optimization

What to Expect

Tesla Energy is growing into hundreds of thousands of sites and multi-gigawatt-hours of deployed capacity. At that scale, a minuscule failure rate is still thousands of real service actions. Every failure mode, data collection failure, and under-reported hardware issue has a price: service visits we don't need, excess warranty cost, and customer pain. Today, fleet health is fragmented: detection, root-cause tracking, fix-efficacy, and "fleet health" metrics lack cohesion across products. Service Engineering Programs is standing up Fleet Health Optimization to own that loop — detect, diagnose, decide, deploy, verify.

This is not a status-tracking role and it is not a data-engineering role. You are a highly technical program manager: you will write the SQL and know our systems cold, then set the agenda for the data engineering and machine learning teams who build the pipelines and detection. You will be the architect of the system that will know where and when any part fails — or is about to — the cost of that failure, and how material and labor capacity constrain solutions. You will own the roadmap and sprint cadence for the program.

What You’ll Do

• Predict time-to-fix, not just time-to-fail — tie the failure forecast to parts availability and crew capacity so we can commit a fix date and hold an SLA to fleet-wide time-to-resolution
• Own Fleet Health Optimization end-to-end — run the Detect→Diagnose→Decide→Deploy→Verify loop as a weekly operating rhythm, hold one issue register with RCA and corrective-action SLAs, and be the single front door that routes emerging signals to the right product vertical
• Do the analysis yourself across many distinct systems — write the SQL and quantify where fault modes, symptoms, and cost joins under-report issues and mislead prioritization
• Make trustworthy operational data step one — treat hardware performance reporting — failure, cost, and part-life data — as the foundation, and architect the genealogy (as-running BOMs; genealogy completeness as change actions flow through systems) so every analysis knows what was built and serviced
• Set the analytical agenda for SEIA data engineering and machine learning — decompose stakeholder asks into engineer-ready work, review SQL and model outputs, and rank what detection is worth building next
• Drive campaign-aware reliability forecasting — fold campaigns, containments, residual risk, and service cost into the forecast so it serves service P&L, not aggregate failure counts
• Own the fix-efficacy and early-detection loops — prove whether a campaign, firmware change, or process fix moved the failure rate, and stand up proactive multi-signal detection ahead of lagging counts
• Partner with vertical leads, Failure Analysis, Factory Quality, Product Lifecycle, and Parts, and own the fleet-health scorecard — time-to-detect, automated-detection share, RCA cycle time, avoided cost, fix-efficacy — for exec forums

What You’ll Bring

• Degree in Engineering, Computer Science, Data Science, Operations Research, or a related field, or equivalent experience
• 5+ years in technical program management, reliability/operations analysis, or a hybrid analyst-TPM role in a hardware and software product environment; 8+ preferred; a former software, data, or reliability engineer who chose to lead programs is a strong fit
• Operational data facility is mandatory — you independently query and reason over service ops systems (tickets, telemetry, hardware lifecycle, cost, contact / 1 month-in-service coding) without waiting for an analyst to hand you a chart
• SQL fluency for original analysis; comfort with GitHub and version control; ability to read a pipeline and sanity-check a machine-learning/detection output before it drives a leadership decision
• Track record translating messy cross-functional problems into scoped engineering workstreams — and closing them with measurable outcomes, not slideware
• Reliability and fleet-health literacy — failure-rate curves, campaign populations, warranty/cost data, and the judgment to challenge a forecast that omits critical modes
• Influence without authority across engineering, analytics, field quality, factory quality, and operations; extreme ownership — you drive problems to closure instead of handing them off