Senior AI Application Engineer - LLM & VLM Enablement
New
A
Axelera AIAI hardware
Work fully remotely from any European country (incl. the UK) you are already in.Full-TimeSenior
Salary not disclosed
Apply NowOpens the employer's application page
Job Details
- Experience
- 5+ years of software engineering experience, with at least 2 years on ML inference, ML systems or ML tooling.
- Required Skills
- PythonPyTorchLLM
Requirements
- Have 5+ years of software engineering experience, including at least 2 years in ML inference, ML systems, or ML tooling.
- Understand transformer inference, including attention variants, KV caching, prefill versus decode, sampling, and tokenization.
- Understand how a VLM combines a vision encoder with a language model.
- Have working knowledge of LLM quantization, including weight-only and weight-activation schemes, calibration, and outlier handling.
- Be able to distinguish quantization loss from an implementation bug.
- Have strong Python and PyTorch skills and experience shipping production-quality software.
- Use AI coding agents daily on real codebases and exercise judgment in verifying or rejecting their output.
- Be able to turn research code and hardware constraints into a practical design.
- Preferred: experience with inference engines or serving stacks such as llama.cpp, vLLM, ONNX Runtime GenAI, or ExecuTorch.
- Preferred: experience with kernel development, tensor compilers such as Triton, TVM, or MLIR, or profiling models on accelerator hardware.
- Preferred: experience with speech (ASR, TTS), agentic applications on embedded devices, ML evaluation or testing harnesses, or tool-calling agents such as MCP servers.
Responsibilities
- Study new LLM and VLM families and architectures, and prototype and profile them on Metis and Europa.
- Determine how models should run on the hardware and productize results as supported Voyager SDK features.
- Create PyTorch references with quantization, weight mappings, and contracts for model inputs, outputs, weights, and persistent state.
- Integrate LLM and VLM serving into the Voyager SDK pipeline builder, including prefill, decode, KV-cache management, streaming text output, and multimodal inputs.
- Measure accuracy against floating-point references with regression and quantization owners, and close or explain each gap.
- Extend the model enablement platform for LLM and VLM bring-up, including hardware-accessible environments, closed-loop runs, and evaluation scoring.
- Recommend Voyager SDK capabilities for GenAI, agents, and speech on edge devices, and identify model zoo demos.
View Full Description & ApplyYou'll be redirected to the employer's site
