Dynamic Approximate Computing Engineer for AI Accelerators

Inria; Rennes; Hybrid; EUR 27,000 - 33,000; Full time

Benefits offered by this job

Teleworking after 6 months

7 weeks annual leave + RTT

Professional equipment

Social, cultural and sports events

Job summary

Inria Rennes Center invites applications for an Engineer F/M to develop energy-efficient AI hardware accelerators using dynamic approximate computing. The role blends hardware design, computer architecture, and co-design with software to push performance and efficiency for edge and data-center AI workloads.

You will model, simulate, and prototype techniques like pruning and mixed-precision arithmetic, and present findings to a research team.

Qualifications

• Master’s degree in Computer Engineering or related fields.
• Strong HW design skills with VHDL/Verilog knowledge.
• Experience with ML frameworks (PyTorch/TensorFlow).
• Knowledge of approximate computing techniques (e.g., pruning, mixed precision).
• Excellent analytical and problem‑solving abilities; focus on energy efficiency.
• Strong communication skills and ability to present research clearly.
• English proficiency (written and spoken).
• Ability to work in a research team and present progress regularly.

Responsibilities

• Investigate and advance energy-efficient AI accelerators using dynamic approximate computing.
• Develop hardware-software co-design methodologies.
• Model, simulate, and prototype with representative AI workloads.
• Systematically evaluate robustness and scalability of proposed approaches.

Skills

HW design

Computer architecture

ML frameworks

Approximate computing

Analytical skills

Communication

English proficiency

Teamwork

Education

Master's degree in Computer Engineering

Job description

Inria Rennes Center invites applications for an Engineer F/M to develop energy-efficient AI hardware accelerators using dynamic approximate computing. The role blends hardware design, computer architecture, and co-design with software to push performance and efficiency for edge and data-center AI workloads.

You will model, simulate, and prototype techniques like pruning and mixed-precision arithmetic, and present findings to a research team.