Research Engineer – Robust Hashing & Representation Algorithms
2 days ago
Cambridge
About the Role We are building advanced algorithmic systems that require highly stable, noise-resilient, and transformation-robust representations. These systems must operate reliably even when inputs vary, compress, distort, or shift across different technical contexts. We are looking for a Research Engineer with strong foundations in signal processing, hashing or encoding algorithms, mathematical modelling, and invariance design, and who is comfortable working with and evaluating modern large language models. You will work at the intersection of algorithms, mathematics, and modern computational models, contributing to representation methods that must remain robust under a wide range of transformations. The work is deeply technical, research- and delivery-driven, and highly applied — without being tied to any single domain. What You Will Do • Develop robust algorithms and representation methods that remain stable under transformations, noise, and perturbations., • Design and analyse hashing, encoding, or similarity algorithms with strong invariance properties., • Apply ideas from signal processing, information theory, and nonlinear transforms to real-world data., • Evaluate behaviour of multiple LLMs (including Qwen-series models) under controlled variations or reparametrisations., • Build experimental frameworks to test algorithmic stability, sensitivity, and discriminative power., • Prototype new algorithmic approaches that generalise across diverse input forms., • Work closely with engineers and researchers to integrate algorithmic insights into larger computational systems., • Contribute to internal theory-building around representation robustness.What You Bring, • Strong foundation in signal processing, transforms, hashing, encoding, or information theory., • Ability to design or mathematically analyse novel algorithms beyond standard machine learning approaches., • Experience with invariance, stability, perturbation analysis, or noise modelling., • Solid mathematical background (linear algebra, spectral methods, applied maths)., • Comfortable running structured experiments with multiple LLMs (Qwen models especially welcome)., • Proficiency in Python (NumPy, SciPy, PyTorch/JAX optional but beneficial)., • Curiosity to explore new algorithmic directions and question assumptions., • Engineers, PhD candidates or postdocs in:, • Signal Processing, • Applied Mathematics, • Information Theory, • Cryptography / Hashing Algorithms, • Electrical Engineering (DSP focus), • Computational Physics, • Computer Science (algorithms, similarity, compression, security) Personal Characteristics • Analytical, rigorous, and detail-oriented, • Comfortable exploring abstract concepts and turning them into applied algorithms, • Approaches problems from first principles, • Enjoys working in a small, focused, research-heavy team, • Thrives in early-stage environments with high autonomy, • Motivated by solving challenging, foundational problems