Vannevar builds agentic AI for 21st-century conflict, pre-trained for national security missions and delivering decision advantage. Backed by General Catalyst and Felicis.
About the role
This team owns the systems that turn user intent into execution - from chat and tasking to persistent agent workflows. You will define how users delegate work to software, how those systems operate on real-world data, and how they are monitored, constrained, and improved over time. This is not a traditional roadmap role. You will take in incomplete, conflicting inputs from users, stakeholders, data, and systems - and decide what we should actually build.
What they're looking for
- Synthesize signal into direction. Take in messy, conflicting inputs and determine what actually matters
- Define how agents behave. Shape primitives like objectives, constraints, trust tiers, and evaluation criteria
- Improve output quality. Drive systematic improvements across models, memory, and workflows
- Own product direction end-to-end. Turn ambiguous problems into concrete bets and drive them through delivery
- Work directly with users. Build deep intuition for real-world workflows and needs
- Translate mission signal into product direction. Take in real-world user signal and translate it into clear product direction, including pushing back when needed
More about this role
Vannevar builds AI systems for the Department of War's most consequential missions. We have 125 deployments across every branch and combatant command spanning our core platform and five distinct products. An an example of how we work, when major combat operations started with Iran, we fielded a new product supporting 24/7 operations and 9,000+ users in three months.
How we build is our advantage: we deploy forward with the people who own the mission. Our engineers, product team, and CTO deploy forward to the point of friction, including visiting units in Ukraine. We focus on embedding directly with operational units supporting great power competition with China, combat operations with Iran, and counter-narcotics missions.
We’re hiring a Product Manager for our Mission Agents team.
This team owns the systems that turn user intent into execution - from chat and tasking to persistent agent workflows. You will define how users delegate work to software, how those systems operate on real-world data, and how they are monitored, constrained, and improved over time.
This is not a traditional roadmap role. You will take in incomplete, conflicting inputs from users, stakeholders, data, and...
Browse similar: Startup jobs · Remote jobs