Embedded AI Intern
Internship · Software · Groningen, The Netherlands
Our AI inference IP runs on an FPGA platform. You bring AI workloads onto it with our tools, benchmark them, and help improve the software around the hardware, with direct guidance from the founders.
What you'll do
- Run AI workloads end to end on our FPGA platform: from trained model through our compiler to execution on the IP
- Quantize models at different precisions with our compiler and validate accuracy against the original model
- Extend the ARM firmware (C, AMD Vitis) that controls the IP over AXI to support new workloads
- Extend our Python compiler flow where a workload requires it
- Automate the compile, build, run, and measure loop; benchmark latency, throughput, and accuracy across workloads and precisions
- Report results and debug issues across the model, compiler, firmware, and hardware boundary
What we look for
- BSc or MSc student in computer science, electrical engineering, or a related field
- Proficiency in Python
- Able to read and modify C code (the firmware side)
- Working knowledge of Linux, shell scripting, and Git
- Understanding of how neural networks execute: layers, weights, quantization
- Systematic debugging habits and clear reporting in English
- Preferably already studying at a university in the Netherlands; we cannot sponsor visas at this stage
Nice to have
- FPGA system experience, including AXI interfaces
- AMD Vitis or Vivado toolchain experience
- RTL design experience (Verilog or SystemVerilog)
- Hands-on model quantization or ONNX experience
- Prior benchmarking or performance measurement work
What you get
- Hands-on work at the hardware-software boundary of an AI chip
- Direct mentorship from the founding team
- Responsibility for clearly scoped, real deliverables
Ximplic is an equal opportunity employer. We welcome applications from people of all backgrounds and select on ability and motivation alone.