AI Research Engineer/Scientist (OpenVINO, NNCF)
ΠΡΡΡ & Π‘ΠΎΠΏΡΠΎΠ²ΠΎΠ΄
ΠΠ»Ρ ΠΌΡΡΡΠ° Ρ ΡΡΠΎΠΉ Π²Π°ΠΊΠ°Π½ΡΠΈΠ΅ΠΉ Π½ΡΠΆΠ΅Π½ Plus
ΠΠΏΠΈΡΠ°Π½ΠΈΠ΅ Π²Π°ΠΊΠ°Π½ΡΠΈΠΈ
TL;DR
AI Research Engineer/Scientist (OpenVINO, NNCF): drive the development of the Neural Network Compression Framework with an accent on state-of-the-art compression algorithms tailored for high-performance neural network inference optimization within the OpenVINO ecosystem. Focus on researching and developing quantization, pruning, and sparsity techniques for optimized inference.
Location: Ireland, Leixlip (Hybrid work model: split time between on-site at site and off-site)
Company
Software Team driving customer value through leadership AI technologies, foundational software stacks, and NPU/GPU IP for client and data center markets.
What you will do
- Join the Neural Network Compression Tools team within OpenVINO Developer Tools.
- Research and develop SOTA compression algorithms for neural network inference optimization.
- Implement optimizations specifically tailored for the OpenVINO ecosystem.
- Collaborate on high-performance inference solutions with hardware awareness.
Requirements
- Master's or PhD in Computer Science, Mathematics, or related field (focus on AI/Deep Learning)
- Experience in Python programming and modern paradigms.
- Proven experience in Deep Learning model optimization (Quantization, Pruning, Sparsity).
- Experience with PyTorch or TensorFlow for training, OpenVINO for inference.
- At least 3 years of software development.
- Spoken and written English: upper-intermediate or advanced
Nice to have
- Hardware awareness (CPU, GPU, NPU) and optimizations for memory-constrained environments.
- Familiarity with Large Language Model architectures (Transformers, Diffusers).
Culture & Benefits
- Hybrid work model eligible.
- Shift 1 (Ireland).
- Equal opportunity employer.
ΠΡΠ΄ΡΡΠ΅ ΠΎΡΡΠΎΡΠΎΠΆΠ½Ρ: Π΅ΡΠ»ΠΈ ΡΠ°Π±ΠΎΡΠΎΠ΄Π°ΡΠ΅Π»Ρ ΠΏΡΠΎΡΠΈΡ Π²ΠΎΠΉΡΠΈ Π² ΠΈΡ ΡΠΈΡΡΠ΅ΠΌΡ, ΠΈΡΠΏΠΎΠ»ΡΠ·ΡΡ iCloud/Google, ΠΏΡΠΈΡΠ»Π°ΡΡ ΠΊΠΎΠ΄/ΠΏΠ°ΡΠΎΠ»Ρ, Π·Π°ΠΏΡΡΡΠΈΡΡ ΠΊΠΎΠ΄/ΠΠ, Π½Π΅ Π΄Π΅Π»Π°ΠΉΡΠ΅ ΡΡΠΎΠ³ΠΎ - ΡΡΠΎ ΠΌΠΎΡΠ΅Π½Π½ΠΈΠΊΠΈ. ΠΠ±ΡΠ·Π°ΡΠ΅Π»ΡΠ½ΠΎ ΠΆΠΌΠΈΡΠ΅ "ΠΠΎΠΆΠ°Π»ΠΎΠ²Π°ΡΡΡΡ" ΠΈΠ»ΠΈ ΠΏΠΈΡΠΈΡΠ΅ Π² ΠΏΠΎΠ΄Π΄Π΅ΡΠΆΠΊΡ. ΠΠΎΠ΄ΡΠΎΠ±Π½Π΅Π΅ Π² Π³Π°ΠΉΠ΄Π΅ β