Staff Product Manager - Applied AI Workflow

The Role

Reporting to the VP of Product Management & Process Excellence, you'll own the production workflows that transform raw satellite imagery into customer-grade insights — designing, automating, and continuously improving the pipelines that sit at the heart of how AiDASH delivers value.

You'll start by going deep on our Vegetation Management Workflow (IVMS) — where complexity is highest and the automation upside is largest. From there, the scope grows to cover Asset Inspection & Monitoring (AIMS) and Climate Risk Intelligence (CRIS).

You won't be designing customer-facing product features. You won't be building the internal platform (that's our Platform PM, your closest peer). You'll be designing the operating model that connects them — the steps, frameworks, and policies that govern how an insight gets produced, who or what handles each step, and where humans stay in the loop.

How you'll make an impact:

• Workflow design across products: Define what the production workflow looks like end-to-end for each product: the sequence of steps, the cohort logic (which customers / geographies / products take which path), the handoffs, and the SLAs
• Step-level frameworks: Author the operating frameworks for individual steps — e.g., the image acquisition framework (when do we re-order? from which vendor? what freshness threshold?), the model QC framework (what's the sampling strategy by model age, terrain, sensor?), and similar for every critical step
• Autonomy and human-in-the-loop policy: Decide where the workflow runs autonomously and where humans intervene. Set and own the confidence thresholds at which model output is trusted enough to drop QC. The technical specifics — sampling strategies, model evaluation methods, HITL mechanics — are owned by a pod of applied AI data scientists and analysts you'll partner closely with. You own the policy decision; they own the underlying technical work that informs it
• Cohort logic and CS alignment: Decide which customers get which workflow flavor. CS and leadership are key stakeholders you'll bring along
• Requirements to Platform PM: Translate workflow design into clear system requirements (e.g., "at step X, capture labels with confidence scores and reviewer ID"). Platform PM owns the system spec; you own that the workflow as designed produces the data and outcomes you need
• KPIs: Own the operating metrics — cost per insight, cycle time, % auto-resolved, manual touches per job, quality against SLA. Track and move them

What Success looks like in 12 months,

• A significant IVMS workflow transformation is shipped, stable in production, and delivering its promised cost and automation impact
• A clear, sequenced plan exists for moving IVMS toward an autonomous-by-default workflow, with human intervention narrowed to a well-defined slice
• Manual-touch volume on IVMS is materially down vs. baseline; the cost-per-insight curve bends
• Every critical workflow step has a metric, a target, and a dashboard. The org can answer "how is the workflow performing this week?" without a Slack thread
• DS, Platform, CS, and GIS Ops consistently align to your decisions without escalation

How You Operate:

• Shape the question before deciding the answer: When handed an ambiguous problem, you don't just solve it as posed — you reframe it, sharpen the metric, and tell us when we're optimizing for the wrong thing
• First-principles process thinker: You can look at a 20-step workflow, ask why step 7 exists, and not lose nuance in the process
• Metric-native: You reach for a measurement before a meeting. You don't ship a workflow without a way to tell whether it's working
• Comfortable being the decider: When cohort logic or automation policy is contested, you make the call and defend it. You don't outsource hard calls upward by default
• Credible with technical peers: DS, Platform PM, and engineering find you a strong partner. You can hold a substantive conversation about model performance, confidence thresholds, and system constraints
• Write well, think in writing: Your primary artifacts are documents — workflow charters, frameworks, decision logs
• Influence without authority: CS, GIS Ops, the applied AI pod, and Platform don't report to you. You move them through clarity, credibility, and shared metrics
• Owner mentality: When the workflow underperforms, you don't say "the model was off" or "execution slipped." You own the outcome and find where the design failed

What we're looking for:

• 9-13 years of total experience, with at least 4-6 years as a Product Manager 
• At least one tour owning an internal, operational, or platform-style product — not exclusively customer-facing feature PM 
• A track record of cross-functional ownership at scale — not just shipping, but changing how the org made decisions in a domain
• Demonstrated ownership of operating KPIs (cost, cycle time, throughput, automation %, quality) under real accountability 

Nice to have:

• Experience working on workflows or products that combine model output with human review, and comfort making decisions that depend on model performance
• Background in geospatial / remote sensing / satellite imagery, utilities, climate-tech, or labeling pipelines
• Experience scaling a function from zero at a Series B–D company in a resource-constrained environment

What this role is not:

• Not a feature PM role: You won't be designing what utilities see in the AiDASH product. That's owned by our IVMS / AIMS / CRIS PMs
• Not a platform or internal-tools PM role: You won't be writing PRDs for the labeling tool, QC dashboard, or routing engine. That's owned by our Platform PM — your peer, not your scope.
• Not an operations role: You won't run the workflow at scale. CS is the operating muscle that runs the workflow you design; GIS Ops is a separate execution function
• Not a process consultant or Op Excellence role: You don't map AS-IS / TO-BE flows and hand them over. You own the workflow as a product — its design, metrics, and evolution — and you're accountable to outcomes
• Not a data science role: You don't build models or design technical model evaluation strategies. You partner with the applied AI pod that does, and you make the policy calls their work informs
• Not customer-facing: You won't be in front of customers directly. The customer voice reaches you through CS and feature PMs
• Not a pure strategy or architect role: You'll be hands-on with frameworks, specs, and decisions — and driving adoption of the workflow

We are proud to be an equal-opportunity employer. We are committed to embracing diversity and inclusion in our hiring practices, and we promote a work environment where everyone, from any race, color, religion, sex, sexual orientation, gender identity, or national origin, can do their best work.