Where the future of AI compute is being defined and built, to unlock new levels of machine intelligence.
March 21
🏢 In-office - Bay Area
Where the future of AI compute is being defined and built, to unlock new levels of machine intelligence.
• Define, participate, and develop critical compiler modules including front-end graph optimizations, additional hardware-aware optimizations, back-end code generators and low-level library (kernel) development. • Use performance-driven methodologies to define a comprehensive set of compiler optimizations needed to enable high-performance AI inference solutions. • Collaborate closely with the AI architecture, algorithms and runtime teams to create and optimize end-to-end performant AI solutions. • Interface with teams building chip-simulators, performance models, assemblers, and disassemblers for EnCharge AI architectures. • Create SDKs to interface custom compiler stack with popular AI frameworks & runtimes. • Work closely with the AI Architecture and FPGA platform teams to jointly optimize the compiler stack and architecture - with the goal of enhancing system level performance for AI applications.
• Masters/Ph.D. in EE/CS with >5 years of industry experience in compiler development and / or chip architectures (preferably AI chip architectures). • Proficiency with C++, Python and Systems programming. • >5 years of experience in compiler and / or library design (in R&D and / or products). • Knowledge of industry-standard (and advanced) tools, graph, and intermediate-representation (IR) formats and methodologies including LLVM, MLIR and TVM. • Familiarity with Tensorflow & PyTorch AI frameworks. • Solid understanding of AI Applications and performance bottlenecks. • Excellent verbal and written communication skills.
• Experience with CI/CD. • Knowledge of the end-to-end runtime stack for AI applications.
Apply Now