2 of 2 Flash Jobs in Cambridge

transformer architectures (e.g., GQA, KV-caching) and best practices Expertise in distributed training techniques Familiarity with optimisation strategies for model inference (e.g., dynamic batching, flash attention, speculative decoding) Proven track record of developing and deploying models at scale Expertise in Python, ML frameworks (TensorFlow, PyTorch), and experience with cloud ...

chip or non‐trivial subsystem. Deep understanding of modern microcontroller and processor design, including real‐time behaviour, deterministic interrupt handling, low‐power modes, flash/SRAM hierarchies, and common peripherals (e.g. SPI, I2C, UART, USB, PIO‐style fabrics, ADC/DAC, timers). Equally strong grasp of broader ...