Research Engineer – Zero-Knowledge Verification

This is a fully remote position, open to applicants in Singapore, +1 more country.

📋 Description

• Facilitate the transition of research prototypes into reliable systems for policymakers.

• Lower the costs associated with proving.

• Bridge the gap between existing proofs and advanced inference behavior.

• Create systems that can be reviewed and approved by discerning technical staff.

• Collaborate with colleagues in hardware security and systems, as well as external cryptographers.

• Publish research to ensure that the methodology remains verifiable by others.

• Validate full inference transcripts from large-scale models end to end and document associated costs.

• Bring open-source prototypes to the same level as proprietary systems.

• Enhance data preparation by using reference implementations as correctness benchmarks for AI-assisted optimization.

• Shape the research agenda for the field.

• Design and evaluate sub-sampling strategies under adversarial models.

• Integrate exact floating-point hardware emulation or transition inference to integer arithmetic with batch-invariant kernels.

• Develop proof circuits for low-precision formats such as NVFP4.

• Explore proof-of-useful-work and proof-of-space methodologies.

• Investigate methods for concealing model architecture through proof recursion.

• Conduct red-team assessments of the constraint system.

⛳️ Requirements

• Experience building tangible projects with modern proof systems like Halo2, Plonky3, STARK-based stacks, or similar systems, OR extensive GPU and ML systems experience with a willingness to learn proof systems.

• Proficient in reasoning about soundness, adversaries, assumptions, and critical arguments.

• Capable of writing code that can be scrutinized and tested by others.

• Open to managing engineering responsibilities, including generalizing across model architectures and addressing numerical discrepancies.

• Comfortable in ambiguous, early-stage environments.

• Strong candidates may have published work in applied cryptography, verifiable computation, or succinct proofs.

• Strong candidates may possess experience with low-precision numerics, including rounding modes and accumulation order.

• Strong candidates may have a background in formal verification.

• Knowledge of lookup arguments, recursion, or folding techniques.

• Experience explaining cryptographic principles to non-specialists.

• Familiarity with using AI systems as research partners on correctness-checkable problems.

• Willing to maintain some overlap in working hours with colleagues across different regions.

🏝️ Benefits

• Competitive benefits and leave policies.

• Flexible and remote work options.

• Visa sponsorship available in the UK for candidates looking to relocate.

• Flexibility in working hours with some overlap across regions.