Senior Product Manager, Agent Builder

DataSnipper•New York, NY

1d•Hybrid

About The Position

DataSnipper is hiring a Senior Product Manager for its Agent Builder team in its New York City office. The role owns Agent Builder in Alwin, DataSnipper’s agentic platform for audit and finance. Agent Builder enables firms to create and manage AI agents using their own methodology, guidance, templates, and prior work, across the agents’ lifecycle from definition through retirement. The position will shape the agent definition model, authoring experience, testing and publishing process, and the systems for observing, versioning, governing, and retiring agents. It combines product strategy and technical decision-making, working closely with a new engineering team in New York, Agent Runtime, Evals, and Document Intelligence teams in Amsterdam, customers, and domain experts. The role will help establish how professionals build, run, and govern production agents in a regulated domain. DataSnipper builds AI solutions for audit and finance professionals, emphasizing explainable, traceable outputs. Its Alwin platform runs cloud-based agents across documents including PDF, Word, and Excel, with an audit trail. The company reports that its platform is used by more than 600,000 professionals across 3,000+ organizations in 175 countries.

Requirements

Experience as a product manager on platform or developer-facing products that other people build on; candidates may come from product or technical PM backgrounds.
Experience shipping LLM or agent features in production, with an understanding of context and retrieval, tool calling, evaluations, human review in the loop, and non-determinism.
Comfort with data models and system design, including definition schemas, versions, immutable records, permissions, and relationships between objects.
Experience with products whose lifecycle customers depend on, including versioning, permissions, release, and deprecation.
Sufficient technical depth to make architecture trade-offs with engineering leads and credibly challenge designs.
Strong discovery discipline, including distinguishing what customers said from what the product manager hoped they said.
Hands-on use of AI tools in personal work, including prototyping.
Clear, concise writing.

Nice To Haves

Knowledge of audit or accounting, from practice or from building products for auditors, finance teams, or other expert users in regulated fields.
Familiarity with how firms validate and govern tools before relying on them, and with quality management standards such as ISQM 1.
Experience with knowledge management or enterprise search, including ingestion, retrieval, document-level permissions, and keeping content fresh.
Experience with low-code or configuration products, workflow designers, rule engines, or DSLs.
Practical agent evaluation experience, including datasets, offline and online experiments, failure classification, and measuring the effect of context on quality.
Experience running design-partner programmes with large enterprise customers.

Responsibilities

Set the product direction and roadmap for Agent Builder, making trade-offs explicit.
Determine the right balance between builder flexibility and guardrails so agents can encode real methodology while remaining trustworthy and consistent.
Define how agents are evaluated on real cases, with results reported using a stated method rather than a headline accuracy number.
Design lifecycle capabilities—including versioning, approvals, rollout, and retirement—as core product features.
Keep customer confidentiality central, ensuring that information learned from one customer’s work does not reach another customer.
Prototype flows before writing specifications, then produce concise specifications and decision memos engineers can use.
Work with forward-deployed engineers, solutions architects, and go-to-market teams to turn enterprise rollout learnings into platform capabilities rather than one-off solutions.
Conduct discovery with people who govern agents at customer firms and adjust plans based on evidence.
Own how firm methodology, guidance, templates, and prior work are captured, kept current, permissioned, and supplied to agents when needed.
Own the agent authoring experience and definition model, including steps, tools, knowledge sources, thresholds, and human-in-the-loop gates.
Define how builders evaluate agents against real cases and how agents are reviewed and approved before publication to the Agent Library.
Enable tracing of live agent activity, including what an agent did and which knowledge it used and why.
Design permissions, change history, approvals, safe version rollouts, and clean retirement of agents and outdated knowledge.