Hani Hamdi
Principal Engineer | AI Accelerators & Solutions
Salary / Daily rate
Paris, France
Freelance
Skills
Languages
About me
Intro
Principal AI Engineer with 10+ years specializing in computer vision and deep learning accelerator architecture. Led teams optimizing neural networks for automotive ADAS, achieving maximum efficiency on embedded platforms. Expert in AI solution architecture, HW/SW integration, and deploying ML models from TensorFlow/ONNX to production systems.
Work experience
Principal Engineer - AI Solution Architect
Untether AI
Jun 2024 - Present
2 years 5 months
Paris, France
• Develop and demonstrate solutions based on Untether AI inference accelerator for LLMs, Vision transformers, CNNs type of architectures • Work directly with key customers and partners in different sectors (ADAS/AD, Data center and edge AI) to understand the project requirements and provide the best solutions • Work closely with sales team to develop and drive opportunities pipeline across different markets • Perform in-depth analysis and optimization to ensure the best performance (throughput, latency, accuracy) out of the Untether AI hardware accelerator for CNNs and LLMs type of architecture (KV cache, tiling, bandwidth bottleneck, etc) • Work with different Business Units to adjust the AI compiler/SDK roadmap based on customer feedback and state of the art analysis
Renesas Electronics
Jan 2019 - Jul 2024
5 years 7 months
Paris, France
Senior Staff Engineer - AI platform Architect
Jan 2022 - Jul 2024
2 years 7 months
Paris, France
Staff Engineer - Computer Vision Technical Lead, AI
Jan 2019 - Jul 2024
5 years 7 months
Responsible of AI segment: • Lead and manage a team of engineers responsible for: o Efficient vision algorithm development o State of the art neural networks benchmark and optimizations to achieve maximum efficiency of runtime and BW in Renesas platforms o PoC development of tools to deploy neural networks from open source frameworks (Caffe, TensorFlow, ONNX, … ) in Renesas deep learning accelerators • Specify HW/SW requirements for next Renesas ADAS generations in AI field including deep learning accelerator architecture (TOPS, layers supported, …) , deployment tools and new standards ( SYCL, ONNX …) • Review and select partners in ADAS field • RFI/RFQ answer
Senior Engineer, AI Solution Architect
Renesas Electronics Corporation
Jul 2017 - Jan 2019
1 year 7 months
Paris Area, France
ADAS Algorithms team, Technical Lead: • Exploring deep learning solutions for automotive problems including object detection, semantic segmentation, for Camera and Lidar sensors • Review competition and supported marketing managers to generate value proposition. ADAS customers, Technical Interface: • Extensive direct experience with customers as algorithms/architecture consultant and cross-functional technical interface for internal software component and tools teams. • Supported several projects from start to end for optimal implementation on Renesas platforms
Valeo
Jul 2013 - Jul 2017
4 years 1 month
Paris Area, France
Embedded Vision Engineer - Driving Assistance Research department
Jan 2014 - Jul 2017
3 years 7 months
Paris Area, France
- Adapt and optimize computer vision and machine learning algorithms on embedded DSP platforms with associated CPU and memory bandwidth limitations - Porting of C/C++ vision algorithms to OpenCL/CUDA and perform benchmarks of different implementations in embedded platforms - Benchmark vision algorithms and deep learning networks (CNN) in embedded platforms (TI, NXP, etc) Automated parking, Autonomous driving, Machine learning, Pedestrian detection,Embedded vision algorithm integration Signal processing, MISRA, C-intrinsic, Memory usage, Cross compilation, Embedded camera, Profiling, etc.
Trainee - Driving Assistance Research department
Jul 2013 - Jan 2014
7 months
Paris, France
- Integration of the Front Collision Warning System “FCWS” in the RVC 3 camera Vehicle detection, Object tracking, OMAP-L138, C674x, Low cost camera, etc
Education
Télécom Paris
Diplôme d'ingénieur
2012 - 2014
3 years