Open role
Member of Technical Staff, Verification & Policy
Own the Logic Scrubber: how a directive is verified against policy before authority is granted.
Team Control plane engineering
Location New York, NY / Remote (US)
Type Full-time
Level Principal MTS
Compensation $280,000 – $350,000 base + meaningful equity + benefits
What we are building
Dipp AI Technologies is on a mission to close the gap between AI’s advanced reasoning capabilities and an enterprise’s ability to trust it with real work: verifiable, cost-disciplined, and in control of its own data.
We build Orcher, the agentic control plane that enforces Human-in-the-Role across seven components, and Dipp Intelligence, our Enterprise Superintelligence solution that compounds on verified execution.
Work directly with the founding team, building for customers from startups to Fortune 100 companies across industries. Help enterprises adopt frontier models, open weights and new compute providers without surrendering authority, economics or data control.
About the role
Before Orcher grants an agent authority to act, the directive is checked against enterprise policy, regulatory constraints, and the boundaries of the requesting role. That check is the Logic Scrubber, and you will own it.
The problem is not writing rules. It is compiling policy from many sources — regulatory regimes, internal control frameworks, contractual obligations — into an evaluation that is deterministic, explainable, and fast enough to sit in the request path.
Explainability is a hard requirement. Every deny must carry a reason a compliance officer can act on, and every allow must carry the evidence that justified it.
What you will do
- Design the policy representation and the compiler that turns authored policy into executable evaluation.
- Build deterministic evaluation with explainable outcomes and stable reason codes.
- Implement conflict resolution when regimes, internal controls, and role scopes disagree.
- Build the simulation and dry-run tooling teams use to test policy before enforcing it.
- Own the performance envelope of evaluation inside the gating path.
- Work with compliance leaders at design partners to encode real control frameworks correctly.
What we look for
- Experience building rules engines, policy systems, query planners, compilers, or type checkers.
- Comfort with formal reasoning: decidability, determinism, and the limits of static analysis.
- Familiarity with OPA/Rego, Cedar, Zanzibar-style models, or an equivalent you have built yourself.
- Ability to produce explanations of machine decisions that non-engineers can act on.
- Strong systems performance instincts and the profiling discipline to back them.
- Willingness to sit with compliance teams and translate their language into code.
Experience
- 10+ years in software engineering with significant depth in language, policy, or evaluation systems.
- Has owned a component that other engineering teams depended on and extended.
- Experience in a regulated domain — finance, healthcare, insurance, defence, or public sector — is a strong plus.
Education
- BS, MS, or PhD in Computer Science or a related field, ideally with exposure to programming languages, logic, or formal methods.
- Equivalent practical experience is accepted; we care about the systems you have shipped.
Compensation and benefits
- Meaningful equity with a standard four-year vest.
- Full medical, dental, and vision coverage.
- 401(k) with company contribution and a hardware budget.
- Time and budget to publish research derived from your work.
How we hire
- 01Intro conversation with the founder.
- 02Design discussion on policy compilation and explainability.
- 03Collaborative exercise on a real evaluation trade-off.
- 04Panel with control-plane and applied-research colleagues.
- 05References and offer.