Research Engineer / Research Scientist, Zero-Knowledge Verification

New
S
SASHAI verification
This is a remote role that can be based in Singapore, the UK, or remotely., Remote team members should be comfortable maintaining some working-hour overlap with colleagues across regions.Full-Time
Salary100,000 - 200,000 USD per year
Apply NowOpens the employer's application page

Job Details

Experience
We are hiring this role at a range of levels, from people early in their research career to people who have shipped proof systems in production. We care more about ability and trajectory than years of experience.

Requirements

  • Have built something real with a modern proof system such as Halo2, Plonky3, STARK-based stacks, or a comparable system, or have deep GPU and ML systems experience and be willing to learn proof systems on the job.
  • Be comfortable reasoning about soundness, adversarial behavior, assumptions, and critical points in an argument.
  • Be able to write code that others will scrutinize and attack.
  • Be willing to handle engineering work such as generalizing across model architectures and resolving numerical mismatches.
  • Be comfortable in ambiguous, early-stage environments where defining the problem is part of the work.
  • Strong candidates may have published work in applied cryptography, verifiable computation, or succinct proofs.
  • Strong candidates may have experience with low-precision numerics, including rounding modes and accumulation order.
  • Strong candidates may have formal verification experience, especially with verifier codebases or protocols.
  • Familiarity with lookup arguments, recursion, or folding schemes is useful.
  • Experience explaining cryptographic guarantees to non-specialists is useful.
  • Experience using AI systems as research collaborators on problems with checkable correctness is useful.

Responsibilities

  • Prove a full inference transcript from a frontier-scale model end to end and report its cost.
  • Bring an open prototype to parity with the best proprietary systems.
  • Improve the data-preparation path that dominates current proving time, using reference implementations as correctness oracles for AI-assisted optimization.
  • Help develop the research agenda, including opportunities for AI-assisted research.
  • Design and analyze sub-sampling schemes, including the fraction of blocks to check under different adversary models.
  • Integrate exact emulation of hardware floating-point behavior or explore production inference using integer arithmetic with batch-invariant kernels.
  • Build proof circuits for low-precision formats such as NVFP4.
  • Investigate compute- and memory-exhaustion schemes such as proof-of-useful-work or proof-of-space.
  • Explore hiding model architecture through techniques such as proof recursion.
  • Red-team the constraint system and publish work for others to check.
View Full Description & ApplyYou'll be redirected to the employer's site
100,000 - 200,000 USD per year
Apply Now