EnCharge AI is looking for an exceptional technical leader who can envision, build, and lead algorithm-hardware codesign efforts for the next generation of AI Inference Hardware. You must have a strong track record of AI algorithm-hardware codesign innovations as well as providing technical leadership to teams in AI Algorithms & Hardware codesign space. This position is a strong growth opportunity, with the possibility of transitioning into further leadership roles in the future.
- Enabling quantization-aware training (QAT) and post-training quantization (PTQ) techniques/packages that work with open-source frameworks, including Tensorflow (TF) and/or PyTorch (PT).
- Driving key innovations that enable quantized AI models to work without loss of accuracy on EnCharge hardware.
- Building Python-based SDKs that enable fully automated quantization schemes for AI models.
- Work with software teams to build and enable AI services & SDKs that exploit Neural Architecture Search (NAS) for model builds – with the goal of designing and discovering EnCharge HW optimized models for customer-specific datasets.
- Drive AI algorithms-hardware roadmap features, project management and schedules.
- Working with clients to understand the specifics of their AI models and to translate those requirements to EnCharge teams working in algorithms, hardware, and software.
- Mentor/lead junior engineers across the company.
- Masters / Ph.D. in EE/CS with >5 years of industry experience in AI algorithms & hardware.
- At least 3-5 years of experience with C++, Python, Tensorflow & PyTorch and >1-2 years of experience in the code structure (and modifying operators) in Tensorflow & PyTorch.
- > 1-2 years of experience with Cuda.
- Deep experience with state-of-the-art neural network topologies in various application domains.
- Knowledge and implementation of advanced QAT and PTQ techniques used to build highly quantized AI models.
- 3-5 years of experience with algorithms-hardware codesign.
- Excellent verbal and communication skills.
- Client engagement experience.
- Knowledge of the end-to-end runtime stack for AI applications including TVM.
- Knowledge of AI Hardware libraries and compiler stacks (including LLVM and MLIR).