> Markdown version of [/jobs/ext/4778703-ai-software-engineer](https://www.wearedevelopers.com/jobs/ext/4778703-ai-software-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Ai Software Engineer - **Company:** Intellectual Capital Resources Limited - **Location:** London, United Kingdom - **Experience:** Experienced - **Contract:** Permanent contract - **Skills:** Abstraction Layers, Artificial Intelligence, C++ (Programming Language), Python (Programming Language), Machine Learning, Systems Integration, Transformer Architectures - **Published:** November 19, 2026 - **Apply:** https://find.jobs/jobs-near-me/apply/ats-redirect/?id=3078400708-2 ## About the Role 3 to 5 years of hands-on, real-world experience in AI/ML engineering or applied AI/ML research (not a graduate profile) Direct experience with model scaling work, diagnosing why ideas do or do not hold at larger scale across the maths, architecture, data and systems layers Working knowledge of several efficient-inference techniques, and familiarity with at least one non-Transformer model family Strong Python fundamentals (C++ a plus), with the systems fluency to work close to the hardware boundary Comfortable working without a fixed spec in an early-stage, research-led environment ## Description Our client, an early-stage AI lab building a brain-inspired computing architecture for ultra-efficient inference, is hiring an AI Software Engineer to scale novel research ideas into real systems, with direct impact on how future AI models run on custom silicon. Focus areas (one or more): Model scaling (taking small-scale proofs of concept to a broad model zoo across sizes, tasks and datasets) Efficient inference (quantisation, low-bit pipelines, KV-cache compression, speculative decoding) Non-Transformer architectures (state-space models, diffusion models, sparse and stateful systems) Kernel- and memory-level optimisation, and hardware-aware ML Evaluation infrastructure (benchmark harnesses, ablation tooling, reference implementations) Systems integration (SDK, compiler, runtime and hardware abstraction layers)