Brendan Lynskey
Staff Engineer (Hw, Sw, Ml, Dsp, Fpga) @Blaize
Signup · Get unlimited contacts
WORK HISTORY
Staff Engineer (Hw, Sw, Ml, Dsp, Fpga) @Blaize
Leeds, GB
Blaize (formerly Thinci) is a startup developing a range of energy-efficient SoCs on leading semiconductor processes, for Machine Learning, Computer Vision, etc.This is a dual role- Systems Engineering (HW): System-on-Module (SoM), and PCIe accelerators - SW Engineering: agentic AI (LangGraph), Python for EDA, Video Analytics, SoM was the 2022 Edge AI and Vision Product of the Year Award SoM also won the 2022 Innovation Award (Hardware & Electronics category) presented by the Sacramento Business Journal.We developed commercial and industrial SoM variants. My role included multi-phase DCDC, PoE, flyback, high-speed interfaces, PI, SI, thermal, layout (HDI), environmental testing and conformal coating. Interfaces include MIPI, PCIe gen3, USB 3.0, LPDDR4. Working from concept through to layout and simulation, bringup, test and EMC pre-compliance.EDA tools include the full Cadence suite (DE, Allegro, PowerSI, SystemSI, PowerDC, OptimizePI, etc.), LTspice/LTpowerCADII, and SIMetrix/SIMPLIS.Other highlights- Analysis of competitor systems- SoM prototype (Xilinx Zynq UltraScale+ MPSoC, Petalinux)- characterised & simulated high-speed serial interfaces (inc. 3D extractions), DDR interfaces, and power-delivery systems (TD and FD modelling & measurements)- Reviewed SoC FC-BGA substrates- Deep-dive into SoC SystemVerilog as part of system debug- Various scripting projects to support testing and EDA flow, e.g:* PowerShell / SKILL utility for Cadence releases * Python utility that produces Power Integrity requirements from raw power-sim data (STFT-based time-series analysis)- VSCode extension (JS/TS)- ML development/benchmarking (Python)- Agentic analytics project (Python, Ollama, LangChain/LangGraph, Cline, Qwen)
EDUCATION
The University of Manchester
M.Sc. (part-time), Low Power Systems Integration
The Open University
Diploma (part-time), Mathematics
University of York
M.Sc., Music Technology (DSP)
The University of Manchester
B.Eng. (Hons), Electronic Engineering
SKILLS
ABOUT BRENDAN LYNSKEY
Motivated, industrious multidisciplinary engineer with >25 years of wide-ranging experience in HW / Power / Thermal, FPGA / SoC, SW / AI / DSP- High-performance HW design: multiphase DCDC converters, PI, SI, Thermal, PCB, DDR, PCIe, USB, measurements, modelling (Cadence, LTC tools, SiMetrix/SIMPLIS)- SW development for Machine Learning applications (Python, agents), with agentic-coding experience- General SW and scripting experience (Python, PowerShell, Git, modern C++, HTML, Tcl, shell, TS,)- DSP, ML, numerical processing (Python/NumPy/SciPy/Jupyter, MATLAB/Simulink)- RTL design and verification (VHDL, Verilog)- FPGA - Xilinx Certified Engineer - many device families up to UltraScale+, AWS EC2 F1- EDA tools for simulation and synthesis (Synopsys, Cadence, Mentor)- Embedded CPU system design (inc. MIPS, Arm, APB/AHB/AXI)- Embedded SW, Embedded Linux (C, U-Boot, Buildroot, LKMs/device drivers)- Audio/Video (professional and consumer standards)- Datacomms (MAC and PHY layers)I\'m experienced in 2 broad areas: HW/FPGA/Power, and SW/ML/DSP.In HW, I often work where digital hardware meets the analogue physics. I’m fascinated by challenging hardware design, and find SI/PI very interesting,(currently using Sigrity). Other areas of particular interest are power delivery for challenging accelerator SoCs, high speed serial links and thermal.In SW, I’ve recently had the chance to work on the development of Machine Learning systems. I had the chance to deploy open-weights LLMs in an agentic-coding application.The application areas I\'ve worked in to date are mostly in AI/ML, HW, SoC Debug, Audio-Video, Datacomms/Networking, DSP and Cryptography.I enjoy using automation to improve the development process.I enjoy mathematics and learning new things. I am a huge fan of AI and LLMs.I\'m a conscientious worker, and value clear communication and good working relationships. I\'m adaptable, and relish a challenge.Personal GitHub page: https://github.com/BrendanJamesLynskey
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.