← appliedintuition / Product Manager - DroneStack
brief / art_XwvzN946YEQ
role
model
anthropic/claude-sonnet-4.6
created
2026-06-01T22:10
Company snapshot
Applied Intuition, founded in 2017 and valued at $15B, builds digital infrastructure for physical AI — spanning simulation, tooling, operating systems, and autonomy stacks for automotive, defense, trucking, construction, mining, and agriculture. Eighteen of the top 20 global automakers and the U.S. military are active customers. The DroneStack team (Acuity ISR+Strike autonomy stack + Axle edge compute hardware) is a defense-focused product line enabling autonomous drone swarms in contested environments; the team has moved past field demonstrations into live deployments and is now transitioning from bespoke program deliveries to scalable multi-customer products. Applied Intuition's engineering reputation is strong in simulation and autonomy tooling; specific internal engineering culture details beyond public signals are not available to confirm.
Team stack
Based on the JD and public signals: C++ and Python are likely core languages for the autonomy stack (standard in robotics/aerospace); ROS or a proprietary middleware layer for drone communication is plausible but unconfirmed. Edge compute (Axle hardware) suggests embedded Linux and real-time OS work. Mission control and planning UIs are likely React or similar web frameworks (based on JD references to operator interfaces). The ISR+Strike stack likely involves computer vision, sensor fusion, and path-planning pipelines — specific frameworks (e.g., OpenCV, PyTorch for perception) are inferred but not confirmed by the JD. CI/CD and simulation-in-the-loop testing are likely given Applied Intuition's core simulation heritage.
Likely questions (10)
| area | question | why |
|---|---|---|
| behavioral | Tell me about a time you had to learn a completely new technical domain from scratch and quickly develop a product strategy in it. How did you structure your learning and what did you ship? | JD explicitly states 'you don't need to arrive as a drone or defense expert — what matters is that you're a fast, rigorous learner.' They want evidence of domain-agnostic ramp-up speed. |
| domain | How would you approach the transition from program-driven, bespoke deliveries to a scalable multi-customer product line? Walk me through the framework you'd use to identify the right abstractions. | The JD calls this out as the central challenge: 'move from delivering bespoke solutions for individual programs to building products that scale across customers, platforms, and theaters.' |
| system_design | Design a mission planning and review interface for a non-technical military operator who needs to plan, execute, and debrief an autonomous drone swarm mission without engineering support in the field. What are the core workflows and what would you deprioritize in v1? | JD specifically calls out 'mission control, mission planning, mission review' interfaces and the requirement that they be 'intuitive enough for non-technical users to operate without engineering support in the field.' |
| behavioral | Describe a product you owned end-to-end — from 0-to-1 research through launch and iteration. What was the hardest prioritization call you made about what NOT to build, and how did you make it? | JD requires 'a track record of shipping products, not just features' and explicitly asks for evidence of hard prioritization calls. |
| domain | How would you conduct primary user research with military operators — people whose workflows are classified, who operate in high-stakes environments, and who may be skeptical of outside product people? What would your first 30 days of research look like? | JD calls out 'conduct primary user research — in the field, at bases, and with international partners' as a core responsibility, and the operator environment is highly specialized. |
| system_design | Applied Intuition's DroneStack runs across multiple airframe vendors and mission sets. How would you design a configuration and interface abstraction layer that lets one software product support heterogeneous hardware platforms without becoming a maintenance nightmare? | JD calls out 'define the abstractions, configurations, and interfaces that turn one-off solutions into repeatable products' as a core deliverable. |
| coding | Walk me through a technically complex system you've worked on — ideally spanning hardware and software or real-time constraints. How did you develop enough depth to make informed architectural tradeoffs with the engineering team? | JD requires 'comfort with deep technical systems' and states 'you need to understand the system well enough to make informed tradeoffs' even without writing code. |
| domain | Build a market thesis for autonomous drone products over the next 3–5 years: which DoD service branches, allied nations, or commercial verticals represent the highest-value expansion opportunities for Applied Intuition, and why? | JD explicitly asks the PM to 'develop a point of view on where the autonomous drone market is heading — across DoD service branches, allied nations, and commercial applications.' |
| behavioral | Tell me about a time you had to align a business development or sales team with an engineering team around a product roadmap where their incentives were in tension. How did you resolve it? | JD positions this PM as 'connective tissue between DroneStack engineering, Applied Intuition's defense business development team, and customers' — cross-functional alignment under tension is a daily reality. |
| culture | Applied Intuition is an in-office, high-intensity environment working on defense products with real operational stakes. How do you think about the ethical dimensions of building autonomous strike capabilities, and how does that fit with your personal values? | Defense autonomy (ISR+Strike) carries significant ethical weight; the company deploys on real platforms with real operators. Culture fit around mission alignment is implicitly critical for this role. |
Talking points
- FAA-certified drone pilot (sUAS Remote Pilot Certificate) with a Private Pilot License — I'm not learning the airspace from a textbook. I understand flight operations, airspace classification, and operator decision-making from direct experience, which gives me a credible foundation to conduct user research with military drone operators and ask the right questions about mission planning and execution workflows.
- I've run the full 0-to-1 product lifecycle twice in the last year as a founder — StreamIO AI (Electron/React desktop + iOS app with real-time HLS streaming, multi-agent orchestration, and cross-platform distribution) and Fintellect AI (RAG pipeline, multi-LLM orchestration, App Store launch). Both involved customer discovery, hard prioritization calls, and shipping to real users without a large team — directly matching the JD's requirement for 'a track record of shipping products, not just features.'
- At Intuit, I drove the ICE Self-Service platform from concept to 675M+ engagements at 50K TPS — a programs-to-products transition at enterprise scale. I identified patterns across 20+ mobile apps and 30+ SKUs, defined the abstractions (DevPortal, GitOps config, ICE Playground), and reduced onboarding from 2–3 weeks to minutes. This is precisely the 'bespoke-to-scalable' transition DroneStack needs to make.
- I built AutoEval — an automated visual evaluation system for robot model training that uses screen capture + multimodal AI (Claude/GPT-4V) to score model outputs (grasp poses, segmentation maps, bounding boxes) against natural-language rubrics, cutting evaluation cycles from 72 hours to ~4 minutes. This demonstrates hands-on comfort with robotics tooling (RViz), computer vision outputs, and real-time systems — the technical substrate of the DroneStack autonomy stack.
- My RL Workbench benchmarks 12 algorithms (PPO, GRPO, DAPO, DPO, and more) across TRL, VeRL, OpenRLHF, and NeMo RL with GPU Docker passthrough and live SSE metric streaming — evidence of the systems-level technical depth the JD requires. Combined with my NeurIPS publication and 20+ years spanning C++ neural nets to production PyTorch platforms, I can engage credibly with Applied Intuition's engineering leads on architecture tradeoffs, not just requirements.