Product + platform
User needs, product claims, APIs, SDKs, service boundaries, release paths, diagnostics, and the evidence required to ship and operate well.
Careers
Work on products where context, execution, and outcome all matter.
The problems we care about cross boundaries: information arrives, context has to be resolved, a decision has to run, and the result has to reach a user, workflow, service, or—when required—a device before the moment passes.
We look for people who can own their part deeply and still follow the failure across interfaces. That includes reading traces, questioning stale assumptions, reproducing bad tail behavior, and staying with the problem until the running system explains what is actually happening.
We do not have posted openings at the moment. The general-interest channel is for engineers and technical leaders whose past work shows this kind of depth and operating judgment.
Areas we work in
Deep work at the points where the operating path changes state.
Data + context
History, state, identity, relationships, freshness, lineage, and the data paths that put the right information where the decision runs.
Decision + automation
Models, rules, optimization, permissions, approvals, workflows, and the control path that turns a decision into a real next step.
Runtime + workload flow
Admission, routing, queueing, scheduling, batching, placement, consistency, capacity, and recovery under real load.
Execution + verification
Software services, workflows, operators, and—when required—devices, followed by a clean read of the resulting business or system state.
Systems research
Experiments that challenge an architectural or physical limit and produce evidence strong enough to change the design.
General interest
Show us a system you had to understand the hard way.
Tell us what was failing, what the first explanation got wrong, which measurements changed your view, what you changed in the system, and how you knew the result improved.
Strong candidates usually have a few scars: race conditions, overload, bad state assumptions, deployment surprises, field failures, misleading model metrics, or integrations that looked fine until the whole process ran together.
Security-clearance privacy
Do not submit clearance identifiers, investigation or adjudication dates, sponsoring agency, compartments, SAP/program details, polygraph status, classified project names, or proof documents.