Malmo, Skåne County, Sweden
Software Engineer with over 15 years of embedded systems development experience, with deep expertise in firmware, OS kernels, device drivers and development of Android based products. I have a strong passion for developing performance critical applications. I am skilled in key AI and GPU acceleration technologies including CUDA, Triton, and PyTorch. Experienced in ML compiler development using LLVM and MLIR. I have a strong foundation in C, C++ and Python
Worked on a Machine Learning SDK focused on converting computational graphs from high-level ML frameworks into the TOSA (Tensor Operator Set Architecture) intermediate representation, and subsequently lowering them to SPIR-V for efficient execution on target hardware. Contributed to development of compiler passes for dialect conversion, graph transformation, and target generation, enabling deployment of ML models across heterogeneous compute platforms.
Worked on development of a Virtual Try-on product involving Stable Diffusion, Control Net and Vision Transformers. Involved in dataset curation, model development, training and fine tuning to generate realistic try-on outputs conditioned on clothing and body pose.
Working with various teams developing Android based consumer electronics products based on Qualcomm Platforms. Main responsibilities include designing Audio SW solutions, developing Audio device driver SW, customizing Audio HAL, BSP bring-up and developing device drivers for various devices.
Carrying out embedded systems related projects as an independent consultant. Major areas of work include: - Android OS source level modifications and porting to new hardware. - FreeBSD and Linux device drivers - Firmware development for Time & Attendance Kiosks - OpenWRT firmware customizations. - Software development for 10 gigE IP cores. - Development of IPTV set-top box software. - Network software optimization using DPDK.