About the role

Responsibilities:

Defines a clear vision and roadmap for AI features or products, ensuring they address real user problems and deliver tangible business outcomes, not merely model deployment.
Strong AI/ML literacy, including understanding of model types, training vs. inference, data quality considerations, limitations, and explainability to drive informed trade-offs.
Data and metrics fluency, including knowledge of data sources, labeling processes, evaluation metrics (accuracy, precision/recall, and hallucination rates), and clear linkage to product KPIs.
Works closely with legal and privacy teams to manage bias, safety, and compliance, and designs fallback experiences when AI outputs are uncertain or incorrect.

PM Core:

Tech Decision-Making: Evaluates tech options using cost, risk, and timeline criteria; can challenge engineering proposals with informed questions; understands API and data trade-offs.
Product Craft and Specs: Produces a clear HLD with architecture context, writes an LLD with data flows and edge cases, and contributes to product strategy discussions with business cases.
Analytics and Data-Driven Decisions: Defines success metrics for features; writes basic SQL or uses BI tools; runs simple A/B test analysis.
Delivery and Stakeholders: Manages backlog independently; runs sprint planning and retros; manages expectations across 2-3 stakeholder groups; prepares exec-ready updates.
OKRs and Roadmap: Writes squad-level OKRs with measurable key results, maintains a quarterly roadmap, partners with designers on user flows, and translates research into requirements.

AI Product Skills:

AI-Assisted Product Craft: Uses SpecKit to generate PRDs and specs with editing; builds clickable prototypes using Figma or AI design tools; creates compelling demo scripts.
AI Tools and Prototyping: Uses Cursor/Claude Code to build simple UI prototypes, evaluates AI tools for squad needs, and writes basic prompts with system instructions and examples.
Agent and AI Design Literacy: Writes product requirements for agent features; makes informed choices between AI interaction patterns; understands latency and hallucination trade-offs.
Prompt Engineering: Designs structured prompts with system instructions, examples, and guardrails; tests across edge cases; implements function-calling and JSON mode.

AI Advanced:

Evals and Model Quality: Defines eval criteria for product features (accuracy, latency, safety); works with engineers to build eval datasets; reviews eval results.
AI Model and Cost Literacy: Tracks model costs per feature; makes model-tier decisions based on use case; defines fine-tuning objectives and success criteria with engineers.
AI Architecture and Governance: Writes requirements for RAG and multi-agent systems; implements safety guardrails in product specs; conducts basic red-teaming; ensures data compliance.

Requirements:

Product Manager - AI